{"utterance_id": "5142-33396-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0002.flac", "task_type": "lexical_detection", "query": "warriors", "prompt": "After the audio ends, extract all occurrences of the word \"warriors\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.87, 1.31, "warriors"]], "response_mode": "offline", "answer": "[{\"word\":\"warriors\",\"start\":0.87,\"end\":1.31}]"} {"utterance_id": "5683-32879-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0014.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.27, 7.89, "fingers"]], "response_mode": "offline", "answer": "[{\"word\":\"fingers\",\"start\":7.27,\"end\":7.89}]"} {"utterance_id": "1284-134647-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.36, 4.19, "diocletian"]], "response_mode": "offline", "answer": "[{\"word\":\"diocletian\",\"start\":3.36,\"end\":4.19}]"} {"utterance_id": "5683-32866-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0000.flac", "task_type": "lexical_detection", "query": "carriage", "prompt": "Process the full audio offline. Find every instance of the word \"carriage\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.54, 1.99, "carriage"]], "response_mode": "offline", "answer": "[{\"word\":\"carriage\",\"start\":1.54,\"end\":1.99}]"} {"utterance_id": "2830-3980-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0038.flac", "task_type": "lexical_detection", "query": "jesus", "prompt": "After the audio ends, extract all occurrences of the word \"jesus\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.15, 4.51, "jesus"]], "response_mode": "offline", "answer": "[{\"word\":\"jesus\",\"start\":4.15,\"end\":4.51}]"} {"utterance_id": "2300-131720-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0014.flac", "task_type": "lexical_detection", "query": "leader", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"leader\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.51, 1.91, "leader"]], "response_mode": "offline", "answer": "[{\"word\":\"leader\",\"start\":1.51,\"end\":1.91}]"} {"utterance_id": "3570-5694-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0008.flac", "task_type": "lexical_detection", "query": "consumer", "prompt": "After the audio ends, extract all occurrences of the word \"consumer\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.47, 5.98, "consumer"]], "response_mode": "offline", "answer": "[{\"word\":\"consumer\",\"start\":5.47,\"end\":5.98}]"} {"utterance_id": "3575-170457-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0054.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.53, 5.5, "miss w"], [6.95, 7.22, "anne's"]], "response_mode": "offline", "answer": "[{\"word\":\"miss w\",\"start\":4.53,\"end\":5.5},{\"word\":\"anne's\",\"start\":6.95,\"end\":7.22}]"} {"utterance_id": "5105-28241-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0018.flac", "task_type": "lexical_detection", "query": "remained", "prompt": "Process the full audio offline. Find every instance of the word \"remained\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.16, 1.63, "remained"]], "response_mode": "offline", "answer": "[{\"word\":\"remained\",\"start\":1.16,\"end\":1.63}]"} {"utterance_id": "4970-29095-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0009.flac", "task_type": "lexical_detection", "query": "habitual", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"habitual\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.75, 4.35, "habitual"]], "response_mode": "offline", "answer": "[{\"word\":\"habitual\",\"start\":3.75,\"end\":4.35}]"} {"utterance_id": "1580-141083-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0044.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.58, 3.01, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":2.58,\"end\":3.01}]"} {"utterance_id": "3570-5696-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0003.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.54, 1.73, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":1.54,\"end\":1.73}]"} {"utterance_id": "7176-92135-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0033.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "You will hear a complete utterance. When it ends, list every instance of a time of day you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.79, 2.13, "evening"]], "response_mode": "offline", "answer": "[{\"word\":\"evening\",\"start\":1.79,\"end\":2.13}]"} {"utterance_id": "7176-92135-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0020.flac", "task_type": "semantic_detection", "query": "city", "prompt": "When the utterance is complete, report all instances of a city name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.98, 2.49, "elsinore"]], "response_mode": "offline", "answer": "[{\"word\":\"elsinore\",\"start\":1.98,\"end\":2.49}]"} {"utterance_id": "1320-122612-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0001.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "After listening to the whole recording, output every instance of a weather term with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 0.79, "dews"], [3.62, 4.1, "mists"]], "response_mode": "offline", "answer": "[{\"word\":\"dews\",\"start\":0.42,\"end\":0.79},{\"word\":\"mists\",\"start\":3.62,\"end\":4.1}]"} {"utterance_id": "2830-3979-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.75, 2.12, "luther's"], [6.69, 7.08, "christ's"]], "response_mode": "offline", "answer": "[{\"word\":\"luther's\",\"start\":1.75,\"end\":2.12},{\"word\":\"christ's\",\"start\":6.69,\"end\":7.08}]"} {"utterance_id": "5105-28240-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0009.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.36, 7.03, "helmsman"]], "response_mode": "offline", "answer": "[{\"word\":\"helmsman\",\"start\":6.36,\"end\":7.03}]"} {"utterance_id": "237-134493-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0004.flac", "task_type": "lexical_detection", "query": "breath", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"breath\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.41, 5.78, "breath"]], "response_mode": "offline", "answer": "[{\"word\":\"breath\",\"start\":5.41,\"end\":5.78}]"} {"utterance_id": "672-122797-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0009.flac", "task_type": "lexical_detection", "query": "met", "prompt": "Once the audio has finished, identify every instance of the word \"met\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.24, 1.56, "met"]], "response_mode": "offline", "answer": "[{\"word\":\"met\",\"start\":1.24,\"end\":1.56}]"} {"utterance_id": "1284-1180-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0006.flac", "task_type": "semantic_counting", "query": "country", "prompt": "Listen to the full audio. Count every occurrence of a country name. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.77, "one"], [3.92, 4.29, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.35,\"end\":2.77},{\"word\":\"two\",\"start\":3.92,\"end\":4.29}]"} {"utterance_id": "1221-135767-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0001.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.83, 6.34, "hester"]], "response_mode": "offline", "answer": "[{\"word\":\"hester\",\"start\":5.83,\"end\":6.34}]"} {"utterance_id": "8230-279154-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0031.flac", "task_type": "lexical_counting", "query": "occurrences", "prompt": "Process the full audio offline. For every occurrence of the word \"occurrences\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.69, 3.19, "one"], [8.7, 9.61, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.69,\"end\":3.19},{\"word\":\"two\",\"start\":8.7,\"end\":9.61}]"} {"utterance_id": "7127-75947-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0021.flac", "task_type": "lexical_detection", "query": "tonnay", "prompt": "After listening to the whole recording, output every instance of the word \"tonnay\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.89, 3.17, "tonnay"]], "response_mode": "offline", "answer": "[{\"word\":\"tonnay\",\"start\":2.89,\"end\":3.17}]"} {"utterance_id": "4446-2273-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0023.flac", "task_type": "lexical_detection", "query": "frightened", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"frightened\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.54, 5.02, "frightened"]], "response_mode": "offline", "answer": "[{\"word\":\"frightened\",\"start\":4.54,\"end\":5.02}]"} {"utterance_id": "7729-102255-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0042.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Process the full audio offline. Find every instance of a city name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.0, 8.56, "lexington"]], "response_mode": "offline", "answer": "[{\"word\":\"lexington\",\"start\":8.0,\"end\":8.56}]"} {"utterance_id": "1089-134691-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0020.flac", "task_type": "lexical_detection", "query": "dedalus", "prompt": "Once the audio has finished, identify every instance of the word \"dedalus\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.25, 3.86, "dedalus"]], "response_mode": "offline", "answer": "[{\"word\":\"dedalus\",\"start\":3.25,\"end\":3.86}]"} {"utterance_id": "4077-13751-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0020.flac", "task_type": "lexical_detection", "query": "press", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"press\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.41, 1.69, "press"]], "response_mode": "offline", "answer": "[{\"word\":\"press\",\"start\":1.41,\"end\":1.69}]"} {"utterance_id": "2094-142345-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0012.flac", "task_type": "lexical_detection", "query": "dinah", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"dinah\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.52, 3.8, "dinah"]], "response_mode": "offline", "answer": "[{\"word\":\"dinah\",\"start\":3.52,\"end\":3.8}]"} {"utterance_id": "4077-13751-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0021.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "After the audio ends, extract all occurrences of a state or province name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.26, 6.76, "missouri"]], "response_mode": "offline", "answer": "[{\"word\":\"missouri\",\"start\":6.26,\"end\":6.76}]"} {"utterance_id": "5683-32866-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0013.flac", "task_type": "lexical_detection", "query": "overhanging", "prompt": "Process the full audio offline. Find every instance of the word \"overhanging\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.75, 3.36, "overhanging"]], "response_mode": "offline", "answer": "[{\"word\":\"overhanging\",\"start\":2.75,\"end\":3.36}]"} {"utterance_id": "6829-68769-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0014.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.91, 1.41, "jailer"]], "response_mode": "offline", "answer": "[{\"word\":\"jailer\",\"start\":0.91,\"end\":1.41}]"} {"utterance_id": "260-123288-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0022.flac", "task_type": "lexical_detection", "query": "lost", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"lost\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.35, 1.99, "lost"]], "response_mode": "offline", "answer": "[{\"word\":\"lost\",\"start\":1.35,\"end\":1.99}]"} {"utterance_id": "5105-28240-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0001.flac", "task_type": "semantic_detection", "query": "food", "prompt": "After the audio ends, extract all occurrences of a food name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[15.26, 15.66, "water"]], "response_mode": "offline", "answer": "[{\"word\":\"water\",\"start\":15.26,\"end\":15.66}]"} {"utterance_id": "908-31957-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0010.flac", "task_type": "lexical_detection", "query": "love", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"love\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.45, 0.97, "love"]], "response_mode": "offline", "answer": "[{\"word\":\"love\",\"start\":0.45,\"end\":0.97}]"} {"utterance_id": "1995-1826-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0008.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Do not answer until the audio ends. Then list each instance of a state or province name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.0, 2.47, "county"]], "response_mode": "offline", "answer": "[{\"word\":\"county\",\"start\":2.0,\"end\":2.47}]"} {"utterance_id": "7021-85628-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0018.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.58, 2.04, "mother's"]], "response_mode": "offline", "answer": "[{\"word\":\"mother's\",\"start\":1.58,\"end\":2.04}]"} {"utterance_id": "4992-23283-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0012.flac", "task_type": "semantic_detection", "query": "money", "prompt": "After the audio ends, extract all occurrences of a money or currency term. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.25, 1.74, "fortune"]], "response_mode": "offline", "answer": "[{\"word\":\"fortune\",\"start\":1.25,\"end\":1.74}]"} {"utterance_id": "6930-75918-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0003.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.2, 12.68, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":12.2,\"end\":12.68}]"} {"utterance_id": "1580-141083-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0012.flac", "task_type": "lexical_detection", "query": "ball", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"ball\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.58, 2.87, "ball"]], "response_mode": "offline", "answer": "[{\"word\":\"ball\",\"start\":2.58,\"end\":2.87}]"} {"utterance_id": "4446-2275-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0005.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Once the audio has finished, identify every instance of a time of day. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.19, 2.47, "morning"]], "response_mode": "offline", "answer": "[{\"word\":\"morning\",\"start\":2.19,\"end\":2.47}]"} {"utterance_id": "4507-16021-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0054.flac", "task_type": "lexical_detection", "query": "science", "prompt": "Once the audio has finished, identify every instance of the word \"science\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.03, 4.13, "science"]], "response_mode": "offline", "answer": "[{\"word\":\"science\",\"start\":3.03,\"end\":4.13}]"} {"utterance_id": "7021-85628-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0003.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[4.03, 4.3, "half"]], "response_mode": "offline", "answer": "[{\"word\":\"half\",\"start\":4.03,\"end\":4.3}]"} {"utterance_id": "1320-122617-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0039.flac", "task_type": "lexical_detection", "query": "desperate", "prompt": "After the audio ends, extract all occurrences of the word \"desperate\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.93, 6.38, "desperate"]], "response_mode": "offline", "answer": "[{\"word\":\"desperate\",\"start\":5.93,\"end\":6.38}]"} {"utterance_id": "7127-75946-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0019.flac", "task_type": "lexical_detection", "query": "suppressed", "prompt": "Process the full audio offline. Find every instance of the word \"suppressed\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.51, 2.17, "suppressed"]], "response_mode": "offline", "answer": "[{\"word\":\"suppressed\",\"start\":1.51,\"end\":2.17}]"} {"utterance_id": "6829-68771-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0011.flac", "task_type": "lexical_detection", "query": "tables", "prompt": "Process the full audio offline. Find every instance of the word \"tables\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.25, 0.7, "tables"]], "response_mode": "offline", "answer": "[{\"word\":\"tables\",\"start\":0.25,\"end\":0.7}]"} {"utterance_id": "908-31957-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0024.flac", "task_type": "lexical_detection", "query": "faith", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"faith\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.64, 7.13, "faith"]], "response_mode": "offline", "answer": "[{\"word\":\"faith\",\"start\":6.64,\"end\":7.13}]"} {"utterance_id": "121-127105-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0001.flac", "task_type": "lexical_detection", "query": "effective", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"effective\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.33, 2.94, "effective"]], "response_mode": "offline", "answer": "[{\"word\":\"effective\",\"start\":2.33,\"end\":2.94}]"} {"utterance_id": "237-126133-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0015.flac", "task_type": "lexical_detection", "query": "furiously", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"furiously\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.74, 5.42, "furiously"]], "response_mode": "offline", "answer": "[{\"word\":\"furiously\",\"start\":4.74,\"end\":5.42}]"} {"utterance_id": "4446-2273-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0018.flac", "task_type": "lexical_detection", "query": "walk", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"walk\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.57, 1.85, "walk"]], "response_mode": "offline", "answer": "[{\"word\":\"walk\",\"start\":1.57,\"end\":1.85}]"} {"utterance_id": "672-122797-0068", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0068.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.66, 2.43, "fir tree"]], "response_mode": "offline", "answer": "[{\"word\":\"fir tree\",\"start\":1.66,\"end\":2.43}]"} {"utterance_id": "5683-32879-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0001.flac", "task_type": "lexical_detection", "query": "though", "prompt": "Process the full audio offline. Find every instance of the word \"though\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.03, 2.21, "though"]], "response_mode": "offline", "answer": "[{\"word\":\"though\",\"start\":2.03,\"end\":2.21}]"} {"utterance_id": "260-123286-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0011.flac", "task_type": "lexical_detection", "query": "weather", "prompt": "Process the full audio offline. Find every instance of the word \"weather\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.23, 1.56, "weather"]], "response_mode": "offline", "answer": "[{\"word\":\"weather\",\"start\":1.23,\"end\":1.56}]"} {"utterance_id": "4446-2275-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0001.flac", "task_type": "lexical_detection", "query": "smiled", "prompt": "After listening to the whole recording, output every instance of the word \"smiled\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.98, 1.49, "smiled"]], "response_mode": "offline", "answer": "[{\"word\":\"smiled\",\"start\":0.98,\"end\":1.49}]"} {"utterance_id": "5142-33396-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0001.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.84, 4.2, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":3.84,\"end\":4.2}]"} {"utterance_id": "2094-142345-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.33, 3.15, "missus poyser"]], "response_mode": "offline", "answer": "[{\"word\":\"missus poyser\",\"start\":2.33,\"end\":3.15}]"} {"utterance_id": "4507-16021-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0028.flac", "task_type": "lexical_detection", "query": "pass", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"pass\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.38, 3.08, "pass"]], "response_mode": "offline", "answer": "[{\"word\":\"pass\",\"start\":2.38,\"end\":3.08}]"} {"utterance_id": "237-134500-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0016.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.26, 2.67, "lindens"]], "response_mode": "offline", "answer": "[{\"word\":\"lindens\",\"start\":2.26,\"end\":2.67}]"} {"utterance_id": "1188-133604-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0038.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the full audio. After it finishes, find every instance of an ordinal number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.31, 3.66, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":3.31,\"end\":3.66}]"} {"utterance_id": "1580-141083-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0029.flac", "task_type": "lexical_counting", "query": "hurried", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"hurried\" (word, start, end). If none, reply [].", "labels": [[3.37, 3.71, "one"], [5.23, 5.71, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.37,\"end\":3.71},{\"word\":\"two\",\"start\":5.23,\"end\":5.71}]"} {"utterance_id": "4446-2275-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0026.flac", "task_type": "lexical_detection", "query": "closed", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"closed\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.52, 0.91, "closed"]], "response_mode": "offline", "answer": "[{\"word\":\"closed\",\"start\":0.52,\"end\":0.91}]"} {"utterance_id": "1320-122612-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0010.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.7, 7.99, "hair"]], "response_mode": "offline", "answer": "[{\"word\":\"hair\",\"start\":7.7,\"end\":7.99}]"} {"utterance_id": "672-122797-0071", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0071.flac", "task_type": "semantic_detection", "query": "date", "prompt": "Do not answer until the audio ends. Then list each instance of a date as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.57, 5.14, "christmas"]], "response_mode": "offline", "answer": "[{\"word\":\"christmas\",\"start\":4.57,\"end\":5.14}]"} {"utterance_id": "908-31957-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0006.flac", "task_type": "lexical_detection", "query": "open", "prompt": "After the audio ends, extract all occurrences of the word \"open\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.32, 0.73, "open"]], "response_mode": "offline", "answer": "[{\"word\":\"open\",\"start\":0.32,\"end\":0.73}]"} {"utterance_id": "8455-210777-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0014.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.43, 2.84, "player"]], "response_mode": "offline", "answer": "[{\"word\":\"player\",\"start\":2.43,\"end\":2.84}]"} {"utterance_id": "4446-2275-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0024.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.87, 1.08, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":0.87,\"end\":1.08}]"} {"utterance_id": "4507-16021-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0035.flac", "task_type": "lexical_counting", "query": "true", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"true\" with word/start/end. If none, reply [].", "labels": [[0.41, 0.78, "one"], [4.03, 4.35, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.41,\"end\":0.78},{\"word\":\"two\",\"start\":4.03,\"end\":4.35}]"} {"utterance_id": "1580-141083-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0039.flac", "task_type": "lexical_detection", "query": "examination", "prompt": "After listening to the whole recording, output every instance of the word \"examination\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.67, 2.44, "examination"]], "response_mode": "offline", "answer": "[{\"word\":\"examination\",\"start\":1.67,\"end\":2.44}]"} {"utterance_id": "6829-68771-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0016.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "After listening to the whole recording, output timed count words for a person's name (one, two, three, ...) as JSON. If none, reply [].", "labels": [[3.43, 3.93, "one"], [5.39, 5.66, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.43,\"end\":3.93},{\"word\":\"two\",\"start\":5.39,\"end\":5.66}]"} {"utterance_id": "1320-122617-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0012.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.5, 3.92, "uncas"]], "response_mode": "offline", "answer": "[{\"word\":\"uncas\",\"start\":3.5,\"end\":3.92}]"} {"utterance_id": "4992-23283-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0005.flac", "task_type": "lexical_counting", "query": "know", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"know\" with word/start/end. If none, reply [].", "labels": [[0.96, 1.16, "one"], [2.71, 2.98, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.96,\"end\":1.16},{\"word\":\"two\",\"start\":2.71,\"end\":2.98}]"} {"utterance_id": "6930-81414-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0008.flac", "task_type": "lexical_detection", "query": "move", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"move\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.17, 2.39, "move"]], "response_mode": "offline", "answer": "[{\"word\":\"move\",\"start\":2.17,\"end\":2.39}]"} {"utterance_id": "1580-141084-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0008.flac", "task_type": "lexical_detection", "query": "situation", "prompt": "After listening to the whole recording, output every instance of the word \"situation\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.79, 5.34, "situation"]], "response_mode": "offline", "answer": "[{\"word\":\"situation\",\"start\":4.79,\"end\":5.34}]"} {"utterance_id": "237-134493-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0006.flac", "task_type": "lexical_detection", "query": "athlete", "prompt": "After listening to the whole recording, output every instance of the word \"athlete\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.81, 2.35, "athlete"]], "response_mode": "offline", "answer": "[{\"word\":\"athlete\",\"start\":1.81,\"end\":2.35}]"} {"utterance_id": "1995-1837-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0015.flac", "task_type": "lexical_detection", "query": "squares", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"squares\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.26, 0.64, "squares"]], "response_mode": "offline", "answer": "[{\"word\":\"squares\",\"start\":0.26,\"end\":0.64}]"} {"utterance_id": "8455-210777-0061", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0061.flac", "task_type": "lexical_detection", "query": "security", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"security\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.96, 7.54, "security"]], "response_mode": "offline", "answer": "[{\"word\":\"security\",\"start\":6.96,\"end\":7.54}]"} {"utterance_id": "6829-68771-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0026.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.36, 1.23, "eliza parsons"]], "response_mode": "offline", "answer": "[{\"word\":\"eliza parsons\",\"start\":0.36,\"end\":1.23}]"} {"utterance_id": "672-122797-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0014.flac", "task_type": "lexical_detection", "query": "already", "prompt": "When the utterance is complete, report all instances of the word \"already\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.75, 1.54, "already"]], "response_mode": "offline", "answer": "[{\"word\":\"already\",\"start\":0.75,\"end\":1.54}]"} {"utterance_id": "7021-79740-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0013.flac", "task_type": "lexical_detection", "query": "quick", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"quick\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.87, 2.27, "quick"]], "response_mode": "offline", "answer": "[{\"word\":\"quick\",\"start\":1.87,\"end\":2.27}]"} {"utterance_id": "260-123286-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0026.flac", "task_type": "lexical_counting", "query": "two", "prompt": "Listen to the full audio. Count every occurrence of the word \"two\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.13, 0.4, "one"], [4.68, 4.95, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.13,\"end\":0.4},{\"word\":\"two\",\"start\":4.68,\"end\":4.95}]"} {"utterance_id": "121-127105-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0003.flac", "task_type": "lexical_detection", "query": "reproach", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"reproach\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.24, 3.87, "reproach"]], "response_mode": "offline", "answer": "[{\"word\":\"reproach\",\"start\":3.24,\"end\":3.87}]"} {"utterance_id": "7729-102255-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0042.flac", "task_type": "lexical_detection", "query": "traveling", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"traveling\" with fields word, start, end (seconds). If none, reply [].", "labels": [[6.68, 7.27, "traveling"]], "response_mode": "offline", "answer": "[{\"word\":\"traveling\",\"start\":6.68,\"end\":7.27}]"} {"utterance_id": "4992-41806-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0016.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.21, 1.8, "mahogany"]], "response_mode": "offline", "answer": "[{\"word\":\"mahogany\",\"start\":1.21,\"end\":1.8}]"} {"utterance_id": "8224-274381-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0000.flac", "task_type": "lexical_detection", "query": "enterprise", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"enterprise\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.37, 3.26, "enterprise"]], "response_mode": "offline", "answer": "[{\"word\":\"enterprise\",\"start\":2.37,\"end\":3.26}]"} {"utterance_id": "5142-33396-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0017.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.68, 0.97, "many"]], "response_mode": "offline", "answer": "[{\"word\":\"many\",\"start\":0.68,\"end\":0.97}]"} {"utterance_id": "5105-28240-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0015.flac", "task_type": "lexical_detection", "query": "stupefied", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"stupefied\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.3, 3.17, "stupefied"]], "response_mode": "offline", "answer": "[{\"word\":\"stupefied\",\"start\":2.3,\"end\":3.17}]"} {"utterance_id": "237-134493-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0017.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.91, 6.36, "farmer"]], "response_mode": "offline", "answer": "[{\"word\":\"farmer\",\"start\":5.91,\"end\":6.36}]"} {"utterance_id": "2830-3980-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0010.flac", "task_type": "lexical_counting", "query": "calls", "prompt": "Listen to the full audio. Count every occurrence of the word \"calls\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.66, 0.93, "one"], [3.29, 3.59, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.66,\"end\":0.93},{\"word\":\"two\",\"start\":3.29,\"end\":3.59}]"} {"utterance_id": "8455-210777-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0026.flac", "task_type": "lexical_detection", "query": "alas", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"alas\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.25, 2.94, "alas"]], "response_mode": "offline", "answer": "[{\"word\":\"alas\",\"start\":2.25,\"end\":2.94}]"} {"utterance_id": "908-157963-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0009.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "After listening to the whole recording, output every instance of a state or province name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.11, 2.67, "har"]], "response_mode": "offline", "answer": "[{\"word\":\"har\",\"start\":2.11,\"end\":2.67}]"} {"utterance_id": "7176-88083-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0019.flac", "task_type": "lexical_detection", "query": "above", "prompt": "Process the full audio offline. Find every instance of the word \"above\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.35, 4.54, "above"]], "response_mode": "offline", "answer": "[{\"word\":\"above\",\"start\":4.35,\"end\":4.54}]"} {"utterance_id": "8463-287645-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0001.flac", "task_type": "lexical_detection", "query": "necessary", "prompt": "After the audio ends, extract all occurrences of the word \"necessary\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.86, 1.55, "necessary"]], "response_mode": "offline", "answer": "[{\"word\":\"necessary\",\"start\":0.86,\"end\":1.55}]"} {"utterance_id": "61-70968-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0015.flac", "task_type": "lexical_detection", "query": "stroller", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"stroller\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.69, 5.18, "stroller"]], "response_mode": "offline", "answer": "[{\"word\":\"stroller\",\"start\":4.69,\"end\":5.18}]"} {"utterance_id": "7176-92135-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0004.flac", "task_type": "lexical_detection", "query": "cannot", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"cannot\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.79, 1.18, "cannot"]], "response_mode": "offline", "answer": "[{\"word\":\"cannot\",\"start\":0.79,\"end\":1.18}]"} {"utterance_id": "8455-210777-0062", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0062.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.43, 2.01, "john bright"]], "response_mode": "offline", "answer": "[{\"word\":\"john bright\",\"start\":1.43,\"end\":2.01}]"} {"utterance_id": "4970-29095-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0036.flac", "task_type": "lexical_detection", "query": "wants", "prompt": "After the audio ends, extract all occurrences of the word \"wants\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.23, 4.75, "wants"]], "response_mode": "offline", "answer": "[{\"word\":\"wants\",\"start\":4.23,\"end\":4.75}]"} {"utterance_id": "2830-3980-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0053.flac", "task_type": "lexical_detection", "query": "paul", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"paul\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.62, 1.86, "paul"]], "response_mode": "offline", "answer": "[{\"word\":\"paul\",\"start\":1.62,\"end\":1.86}]"} {"utterance_id": "6829-68769-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0015.flac", "task_type": "lexical_counting", "query": "i'm", "prompt": "Once the audio has finished, list each occurrence of the word \"i'm\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[1.24, 1.52, "one"], [2.86, 2.98, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.24,\"end\":1.52},{\"word\":\"two\",\"start\":2.86,\"end\":2.98}]"} {"utterance_id": "7176-88083-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0016.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an animal name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.09, 1.55, "hawk"]], "response_mode": "offline", "answer": "[{\"word\":\"hawk\",\"start\":1.09,\"end\":1.55}]"} {"utterance_id": "4970-29095-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0013.flac", "task_type": "lexical_detection", "query": "practice", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"practice\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.82, 1.33, "practice"]], "response_mode": "offline", "answer": "[{\"word\":\"practice\",\"start\":0.82,\"end\":1.33}]"} {"utterance_id": "61-70968-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0027.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "When the utterance is complete, report how a profession accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.97, 1.39, "one"], [2.63, 3.03, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.97,\"end\":1.39},{\"word\":\"two\",\"start\":2.63,\"end\":3.03}]"} {"utterance_id": "672-122797-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0030.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "When the utterance is complete, report all instances of a plant name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.43, 2.0, "trees"]], "response_mode": "offline", "answer": "[{\"word\":\"trees\",\"start\":1.43,\"end\":2.0}]"} {"utterance_id": "7127-75947-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0028.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the full audio. After it finishes, find every instance of a plant name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.12, 2.85, "reed grass"]], "response_mode": "offline", "answer": "[{\"word\":\"reed grass\",\"start\":2.12,\"end\":2.85}]"} {"utterance_id": "8224-274384-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0010.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an organization name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.84, 3.81, "administration"], [5.87, 6.51, "parliaments"]], "response_mode": "offline", "answer": "[{\"word\":\"administration\",\"start\":2.84,\"end\":3.81},{\"word\":\"parliaments\",\"start\":5.87,\"end\":6.51}]"} {"utterance_id": "2830-3980-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0037.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.65, 6.03, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":5.65,\"end\":6.03}]"} {"utterance_id": "3729-6852-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0036.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "Listen to the entire clip, then report the count sequence for a profession as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.62, 0.9, "one"], [4.28, 4.93, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.62,\"end\":0.9},{\"word\":\"two\",\"start\":4.28,\"end\":4.93}]"} {"utterance_id": "6930-81414-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0007.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.02, 2.46, "wrist"]], "response_mode": "offline", "answer": "[{\"word\":\"wrist\",\"start\":2.02,\"end\":2.46}]"} {"utterance_id": "1580-141083-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0010.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen carefully to the entire clip, then report every instance of a food name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.91, 1.32, "brandy"]], "response_mode": "offline", "answer": "[{\"word\":\"brandy\",\"start\":0.91,\"end\":1.32}]"} {"utterance_id": "8224-274381-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0007.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[4.86, 5.45, "general"]], "response_mode": "offline", "answer": "[{\"word\":\"general\",\"start\":4.86,\"end\":5.45}]"} {"utterance_id": "2961-961-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0009.flac", "task_type": "lexical_detection", "query": "heard", "prompt": "After the audio ends, extract all occurrences of the word \"heard\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.37, 4.67, "heard"]], "response_mode": "offline", "answer": "[{\"word\":\"heard\",\"start\":4.37,\"end\":4.67}]"} {"utterance_id": "1995-1837-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0001.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Once the audio has finished, identify every instance of a plant name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.35, 4.81, "cotton"]], "response_mode": "offline", "answer": "[{\"word\":\"cotton\",\"start\":4.35,\"end\":4.81}]"} {"utterance_id": "4992-41806-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0000.flac", "task_type": "lexical_counting", "query": "kitchen", "prompt": "When the utterance is complete, report how the word \"kitchen\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.98, 2.34, "one"], [8.26, 8.69, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.98,\"end\":2.34},{\"word\":\"two\",\"start\":8.26,\"end\":8.69}]"} {"utterance_id": "3575-170457-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0026.flac", "task_type": "semantic_counting", "query": "ordinal", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for an ordinal number with word/start/end. If none, reply [].", "labels": [[3.58, 3.9, "one"], [14.42, 14.75, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.58,\"end\":3.9},{\"word\":\"two\",\"start\":14.42,\"end\":14.75}]"} {"utterance_id": "1995-1826-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0004.flac", "task_type": "lexical_detection", "query": "learn", "prompt": "After listening to the whole recording, output every instance of the word \"learn\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.6, 0.93, "learn"]], "response_mode": "offline", "answer": "[{\"word\":\"learn\",\"start\":0.6,\"end\":0.93}]"} {"utterance_id": "4992-23283-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0015.flac", "task_type": "lexical_detection", "query": "inclinations", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"inclinations\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.15, 3.21, "inclinations"]], "response_mode": "offline", "answer": "[{\"word\":\"inclinations\",\"start\":2.15,\"end\":3.21}]"} {"utterance_id": "1089-134686-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0024.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.62, 7.67, "jesus christ"]], "response_mode": "offline", "answer": "[{\"word\":\"jesus christ\",\"start\":6.62,\"end\":7.67}]"} {"utterance_id": "6829-68769-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0010.flac", "task_type": "lexical_detection", "query": "kenneth", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"kenneth\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.93, 2.33, "kenneth"]], "response_mode": "offline", "answer": "[{\"word\":\"kenneth\",\"start\":1.93,\"end\":2.33}]"} {"utterance_id": "672-122797-0064", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0064.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.79, 2.31, "mice"], [11.32, 11.66, "mice"]], "response_mode": "offline", "answer": "[{\"word\":\"mice\",\"start\":1.79,\"end\":2.31},{\"word\":\"mice\",\"start\":11.32,\"end\":11.66}]"} {"utterance_id": "1089-134686-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0022.flac", "task_type": "lexical_counting", "query": "promises", "prompt": "Listen to the entire clip, then report the count sequence for the word \"promises\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.91, 3.49, "one"], [7.22, 7.86, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.91,\"end\":3.49},{\"word\":\"two\",\"start\":7.22,\"end\":7.86}]"} {"utterance_id": "1580-141084-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.37, 1.84, "watson"]], "response_mode": "offline", "answer": "[{\"word\":\"watson\",\"start\":1.37,\"end\":1.84}]"} {"utterance_id": "2830-3980-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0005.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.43, 2.74, "few"]], "response_mode": "offline", "answer": "[{\"word\":\"few\",\"start\":2.43,\"end\":2.74}]"} {"utterance_id": "61-70970-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0032.flac", "task_type": "lexical_detection", "query": "suspicions", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"suspicions\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.49, 2.11, "suspicions"]], "response_mode": "offline", "answer": "[{\"word\":\"suspicions\",\"start\":1.49,\"end\":2.11}]"} {"utterance_id": "61-70970-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0022.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.32, 0.78, "robin"]], "response_mode": "offline", "answer": "[{\"word\":\"robin\",\"start\":0.32,\"end\":0.78}]"} {"utterance_id": "5639-40744-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0024.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.48, 0.72, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.48,\"end\":0.72}]"} {"utterance_id": "6829-68769-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0045.flac", "task_type": "semantic_detection", "query": "money", "prompt": "After listening to the whole recording, output every instance of a money or currency term with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.92, 1.25, "costs"]], "response_mode": "offline", "answer": "[{\"word\":\"costs\",\"start\":0.92,\"end\":1.25}]"} {"utterance_id": "7021-79759-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79759/7021-79759-0005.flac", "task_type": "lexical_counting", "query": "pass", "prompt": "Listen to the full audio. Count every occurrence of the word \"pass\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.59, 8.92, "one"], [11.41, 11.69, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":8.59,\"end\":8.92},{\"word\":\"two\",\"start\":11.41,\"end\":11.69}]"} {"utterance_id": "8224-274381-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0015.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Process the full audio offline. Find every instance of an organization name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.97, 3.4, "army"]], "response_mode": "offline", "answer": "[{\"word\":\"army\",\"start\":2.97,\"end\":3.4}]"} {"utterance_id": "7127-75947-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0011.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen carefully to the entire clip, then report every instance of a time of day as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.11, 2.41, "evening"]], "response_mode": "offline", "answer": "[{\"word\":\"evening\",\"start\":2.11,\"end\":2.41}]"} {"utterance_id": "121-121726-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0003.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.39, 5.87, "grass"]], "response_mode": "offline", "answer": "[{\"word\":\"grass\",\"start\":5.39,\"end\":5.87}]"} {"utterance_id": "8463-294828-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0027.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Once the audio has finished, identify every instance of a country name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.18, 5.73, "france"]], "response_mode": "offline", "answer": "[{\"word\":\"france\",\"start\":5.18,\"end\":5.73}]"} {"utterance_id": "61-70970-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0035.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.51, 1.07, "warrenton"], [3.1, 3.49, "robin"], [4.54, 4.95, "geoffrey's"]], "response_mode": "offline", "answer": "[{\"word\":\"warrenton\",\"start\":0.51,\"end\":1.07},{\"word\":\"robin\",\"start\":3.1,\"end\":3.49},{\"word\":\"geoffrey's\",\"start\":4.54,\"end\":4.95}]"} {"utterance_id": "1580-141083-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0050.flac", "task_type": "lexical_detection", "query": "really", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"really\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.58, 0.86, "really"]], "response_mode": "offline", "answer": "[{\"word\":\"really\",\"start\":0.58,\"end\":0.86}]"} {"utterance_id": "2300-131720-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0021.flac", "task_type": "lexical_counting", "query": "construction", "prompt": "Process the full audio offline. For every occurrence of the word \"construction\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.29, 3.93, "one"], [5.6, 6.23, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.29,\"end\":3.93},{\"word\":\"two\",\"start\":5.6,\"end\":6.23}]"} {"utterance_id": "5105-28241-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0019.flac", "task_type": "lexical_detection", "query": "disappointment", "prompt": "After the audio ends, extract all occurrences of the word \"disappointment\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.55, 4.32, "disappointment"]], "response_mode": "offline", "answer": "[{\"word\":\"disappointment\",\"start\":3.55,\"end\":4.32}]"} {"utterance_id": "8555-284447-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0015.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Listen to the full audio. Count every occurrence of a body part. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.54, 10.97, "one"], [12.92, 13.48, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":10.54,\"end\":10.97},{\"word\":\"two\",\"start\":12.92,\"end\":13.48}]"} {"utterance_id": "7021-85628-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.54, 2.01, "anders"]], "response_mode": "offline", "answer": "[{\"word\":\"anders\",\"start\":1.54,\"end\":2.01}]"} {"utterance_id": "1320-122617-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0012.flac", "task_type": "lexical_detection", "query": "captive", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"captive\" with fields word, start, end (seconds). If none, reply [].", "labels": [[6.4, 7.12, "captive"]], "response_mode": "offline", "answer": "[{\"word\":\"captive\",\"start\":6.4,\"end\":7.12}]"} {"utterance_id": "6829-68771-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0031.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.52, 1.81, "maid's"]], "response_mode": "offline", "answer": "[{\"word\":\"maid's\",\"start\":1.52,\"end\":1.81}]"} {"utterance_id": "908-31957-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0000.flac", "task_type": "lexical_detection", "query": "word", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"word\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.3, 1.91, "word"]], "response_mode": "offline", "answer": "[{\"word\":\"word\",\"start\":1.3,\"end\":1.91}]"} {"utterance_id": "6829-68769-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0052.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen to the full audio. After it finishes, find every instance of a money or currency term. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.84, 2.15, "check"]], "response_mode": "offline", "answer": "[{\"word\":\"check\",\"start\":1.84,\"end\":2.15}]"} {"utterance_id": "7176-88083-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0011.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[3.58, 3.96, "tail"], [6.56, 6.91, "gullet"]], "response_mode": "offline", "answer": "[{\"word\":\"tail\",\"start\":3.58,\"end\":3.96},{\"word\":\"gullet\",\"start\":6.56,\"end\":6.91}]"} {"utterance_id": "1188-133604-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0004.flac", "task_type": "lexical_detection", "query": "drops", "prompt": "Once the audio has finished, identify every instance of the word \"drops\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.72, 6.2, "drops"]], "response_mode": "offline", "answer": "[{\"word\":\"drops\",\"start\":5.72,\"end\":6.2}]"} {"utterance_id": "4992-41806-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0014.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Listen to the entire clip, then report the count sequence for a number as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.24, 10.46, "one"], [12.68, 13.41, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":10.24,\"end\":10.46},{\"word\":\"two\",\"start\":12.68,\"end\":13.41}]"} {"utterance_id": "260-123288-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0026.flac", "task_type": "semantic_detection", "query": "color", "prompt": "You will hear a complete utterance. When it ends, list every instance of a color name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.88, 9.24, "blue"]], "response_mode": "offline", "answer": "[{\"word\":\"blue\",\"start\":8.88,\"end\":9.24}]"} {"utterance_id": "260-123286-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0009.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.05, 4.48, "lips"]], "response_mode": "offline", "answer": "[{\"word\":\"lips\",\"start\":4.05,\"end\":4.48}]"} {"utterance_id": "4970-29093-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0020.flac", "task_type": "lexical_detection", "query": "map", "prompt": "Process the full audio offline. Find every instance of the word \"map\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.83, 5.1, "map"]], "response_mode": "offline", "answer": "[{\"word\":\"map\",\"start\":4.83,\"end\":5.1}]"} {"utterance_id": "3575-170457-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0047.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Do not answer until the audio ends. Then list each instance of an animal name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.39, 0.72, "tabby"]], "response_mode": "offline", "answer": "[{\"word\":\"tabby\",\"start\":0.39,\"end\":0.72}]"} {"utterance_id": "2961-961-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0003.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.52, 1.94, "timaeus"]], "response_mode": "offline", "answer": "[{\"word\":\"timaeus\",\"start\":1.52,\"end\":1.94}]"} {"utterance_id": "1188-133604-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0007.flac", "task_type": "semantic_counting", "query": "color", "prompt": "Process the full audio offline. For every occurrence of a color name, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.59, 0.94, "one"], [4.35, 4.77, "two"], [7.23, 7.87, "three"], [11.16, 11.68, "four"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.59,\"end\":0.94},{\"word\":\"two\",\"start\":4.35,\"end\":4.77},{\"word\":\"three\",\"start\":7.23,\"end\":7.87},{\"word\":\"four\",\"start\":11.16,\"end\":11.68}]"} {"utterance_id": "3575-170457-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0026.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[15.48, 16.22, "c b"]], "response_mode": "offline", "answer": "[{\"word\":\"c b\",\"start\":15.48,\"end\":16.22}]"} {"utterance_id": "237-134500-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0001.flac", "task_type": "lexical_detection", "query": "sighed", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"sighed\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.94, 1.62, "sighed"]], "response_mode": "offline", "answer": "[{\"word\":\"sighed\",\"start\":0.94,\"end\":1.62}]"} {"utterance_id": "3575-170457-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0028.flac", "task_type": "semantic_detection", "query": "date", "prompt": "Listen to the full audio. After it finishes, find every instance of a date. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.29, 3.43, "march twenty second eighteen thirty seven"]], "response_mode": "offline", "answer": "[{\"word\":\"march twenty second eighteen thirty seven\",\"start\":1.29,\"end\":3.43}]"} {"utterance_id": "1995-1836-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0011.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen carefully to the entire clip, then report every instance of a family relation as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.41, 4.0, "sister's"]], "response_mode": "offline", "answer": "[{\"word\":\"sister's\",\"start\":3.41,\"end\":4.0}]"} {"utterance_id": "61-70968-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0022.flac", "task_type": "lexical_detection", "query": "apprentice", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"apprentice\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.04, 4.54, "apprentice"]], "response_mode": "offline", "answer": "[{\"word\":\"apprentice\",\"start\":4.04,\"end\":4.54}]"} {"utterance_id": "4077-13751-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0018.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.19, 8.74, "legate"]], "response_mode": "offline", "answer": "[{\"word\":\"legate\",\"start\":8.19,\"end\":8.74}]"} {"utterance_id": "672-122797-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0010.flac", "task_type": "lexical_detection", "query": "rejoice", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"rejoice\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.5, 1.35, "rejoice"]], "response_mode": "offline", "answer": "[{\"word\":\"rejoice\",\"start\":0.5,\"end\":1.35}]"} {"utterance_id": "4507-16021-0058", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0058.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.41, 2.84, "wing"]], "response_mode": "offline", "answer": "[{\"word\":\"wing\",\"start\":2.41,\"end\":2.84}]"} {"utterance_id": "5639-40744-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0037.flac", "task_type": "semantic_counting", "query": "kinship", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a family relation: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[1.02, 1.35, "one"], [4.16, 4.63, "two"], [7.36, 7.73, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.02,\"end\":1.35},{\"word\":\"two\",\"start\":4.16,\"end\":4.63},{\"word\":\"three\",\"start\":7.36,\"end\":7.73}]"} {"utterance_id": "3575-170457-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0044.flac", "task_type": "lexical_detection", "query": "fatality", "prompt": "Once the audio has finished, identify every instance of the word \"fatality\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[7.59, 8.17, "fatality"]], "response_mode": "offline", "answer": "[{\"word\":\"fatality\",\"start\":7.59,\"end\":8.17}]"} {"utterance_id": "1221-135767-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0004.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After the audio ends, extract all occurrences of a color name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.49, 7.04, "tints"], [14.62, 15.07, "brown"], [18.2, 18.74, "black"]], "response_mode": "offline", "answer": "[{\"word\":\"tints\",\"start\":6.49,\"end\":7.04},{\"word\":\"brown\",\"start\":14.62,\"end\":15.07},{\"word\":\"black\",\"start\":18.2,\"end\":18.74}]"} {"utterance_id": "8224-274381-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0003.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Listen carefully to the entire clip, then report every instance of a country name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.94, 12.69, "scotland"]], "response_mode": "offline", "answer": "[{\"word\":\"scotland\",\"start\":11.94,\"end\":12.69}]"} {"utterance_id": "8224-274381-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0002.flac", "task_type": "semantic_counting", "query": "ordinal", "prompt": "After the audio ends, tally every instance of an ordinal number. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.67, 1.36, "one"], [14.72, 15.26, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.67,\"end\":1.36},{\"word\":\"two\",\"start\":14.72,\"end\":15.26}]"} {"utterance_id": "4446-2273-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0002.flac", "task_type": "lexical_detection", "query": "fancy", "prompt": "After listening to the whole recording, output every instance of the word \"fancy\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.43, 3.01, "fancy"]], "response_mode": "offline", "answer": "[{\"word\":\"fancy\",\"start\":2.43,\"end\":3.01}]"} {"utterance_id": "8230-279154-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0028.flac", "task_type": "lexical_detection", "query": "hardly", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"hardly\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.0, 8.43, "hardly"]], "response_mode": "offline", "answer": "[{\"word\":\"hardly\",\"start\":8.0,\"end\":8.43}]"} {"utterance_id": "2094-142345-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0007.flac", "task_type": "lexical_detection", "query": "plain", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"plain\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.61, 1.9, "plain"]], "response_mode": "offline", "answer": "[{\"word\":\"plain\",\"start\":1.61,\"end\":1.9}]"} {"utterance_id": "1580-141083-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0008.flac", "task_type": "lexical_detection", "query": "proof", "prompt": "Process the full audio offline. Find every instance of the word \"proof\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.24, 0.61, "proof"]], "response_mode": "offline", "answer": "[{\"word\":\"proof\",\"start\":0.24,\"end\":0.61}]"} {"utterance_id": "121-121726-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0006.flac", "task_type": "lexical_detection", "query": "cause", "prompt": "Process the full audio offline. Find every instance of the word \"cause\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.14, 2.58, "cause"]], "response_mode": "offline", "answer": "[{\"word\":\"cause\",\"start\":2.14,\"end\":2.58}]"} {"utterance_id": "8455-210777-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0028.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a family relation with fields word, start, end (seconds). If none, reply [].", "labels": [[2.16, 2.68, "husband"]], "response_mode": "offline", "answer": "[{\"word\":\"husband\",\"start\":2.16,\"end\":2.68}]"} {"utterance_id": "672-122797-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0051.flac", "task_type": "lexical_detection", "query": "means", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"means\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.88, 2.46, "means"]], "response_mode": "offline", "answer": "[{\"word\":\"means\",\"start\":1.88,\"end\":2.46}]"} {"utterance_id": "6829-68771-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0017.flac", "task_type": "lexical_detection", "query": "led", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"led\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.98, 1.18, "led"]], "response_mode": "offline", "answer": "[{\"word\":\"led\",\"start\":0.98,\"end\":1.18}]"} {"utterance_id": "4446-2273-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0017.flac", "task_type": "lexical_detection", "query": "young", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"young\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.63, 1.91, "young"]], "response_mode": "offline", "answer": "[{\"word\":\"young\",\"start\":1.63,\"end\":1.91}]"} {"utterance_id": "8463-294828-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0003.flac", "task_type": "lexical_detection", "query": "botanical", "prompt": "After the audio ends, extract all occurrences of the word \"botanical\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.6, 6.21, "botanical"]], "response_mode": "offline", "answer": "[{\"word\":\"botanical\",\"start\":5.6,\"end\":6.21}]"} {"utterance_id": "7176-92135-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0045.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 1.07, "lord john"]], "response_mode": "offline", "answer": "[{\"word\":\"lord john\",\"start\":0.42,\"end\":1.07}]"} {"utterance_id": "1580-141083-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0042.flac", "task_type": "lexical_detection", "query": "working", "prompt": "After the audio ends, extract all occurrences of the word \"working\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.98, 3.44, "working"]], "response_mode": "offline", "answer": "[{\"word\":\"working\",\"start\":2.98,\"end\":3.44}]"} {"utterance_id": "4992-41797-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0014.flac", "task_type": "lexical_detection", "query": "pictures", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"pictures\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.38, 5.95, "pictures"]], "response_mode": "offline", "answer": "[{\"word\":\"pictures\",\"start\":5.38,\"end\":5.95}]"} {"utterance_id": "4446-2275-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0021.flac", "task_type": "lexical_detection", "query": "whispered", "prompt": "After listening to the whole recording, output every instance of the word \"whispered\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.88, 2.33, "whispered"]], "response_mode": "offline", "answer": "[{\"word\":\"whispered\",\"start\":1.88,\"end\":2.33}]"} {"utterance_id": "3729-6852-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0024.flac", "task_type": "lexical_detection", "query": "crowd", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"crowd\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.92, 1.32, "crowd"]], "response_mode": "offline", "answer": "[{\"word\":\"crowd\",\"start\":0.92,\"end\":1.32}]"} {"utterance_id": "4970-29093-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0020.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "After the audio ends, extract all occurrences of a state or province name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.18, 1.64, "missouri"]], "response_mode": "offline", "answer": "[{\"word\":\"missouri\",\"start\":1.18,\"end\":1.64}]"} {"utterance_id": "2830-3980-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0029.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.82, 5.0, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.82,\"end\":5.0}]"} {"utterance_id": "7021-79740-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0001.flac", "task_type": "lexical_detection", "query": "jane", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"jane\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.26, 5.79, "jane"]], "response_mode": "offline", "answer": "[{\"word\":\"jane\",\"start\":5.26,\"end\":5.79}]"} {"utterance_id": "4446-2275-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.86, "bartley"]], "response_mode": "offline", "answer": "[{\"word\":\"bartley\",\"start\":0.53,\"end\":0.86}]"} {"utterance_id": "1580-141084-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0021.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen carefully to the entire clip, then report every instance of a color name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.04, 2.48, "black"]], "response_mode": "offline", "answer": "[{\"word\":\"black\",\"start\":2.04,\"end\":2.48}]"} {"utterance_id": "260-123440-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0012.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[2.11, 2.48, "heads"]], "response_mode": "offline", "answer": "[{\"word\":\"heads\",\"start\":2.11,\"end\":2.48}]"} {"utterance_id": "1284-1180-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0021.flac", "task_type": "lexical_detection", "query": "brains", "prompt": "Once the audio has finished, identify every instance of the word \"brains\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.66, 4.16, "brains"]], "response_mode": "offline", "answer": "[{\"word\":\"brains\",\"start\":3.66,\"end\":4.16}]"} {"utterance_id": "8230-279154-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0034.flac", "task_type": "lexical_detection", "query": "occurs", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"occurs\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.97, 2.48, "occurs"]], "response_mode": "offline", "answer": "[{\"word\":\"occurs\",\"start\":1.97,\"end\":2.48}]"} {"utterance_id": "1089-134686-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0013.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.06, 7.52, "knight"]], "response_mode": "offline", "answer": "[{\"word\":\"knight\",\"start\":7.06,\"end\":7.52}]"} {"utterance_id": "260-123286-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0026.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "When the utterance is complete, report all instances of an animal name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.4, 0.94, "monsters"], [4.95, 5.67, "reptiles"]], "response_mode": "offline", "answer": "[{\"word\":\"monsters\",\"start\":0.4,\"end\":0.94},{\"word\":\"reptiles\",\"start\":4.95,\"end\":5.67}]"} {"utterance_id": "4992-41797-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0005.flac", "task_type": "lexical_detection", "query": "ain't", "prompt": "After the audio ends, extract all occurrences of the word \"ain't\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.31, 1.48, "ain't"]], "response_mode": "offline", "answer": "[{\"word\":\"ain't\",\"start\":1.31,\"end\":1.48}]"} {"utterance_id": "1995-1826-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0010.flac", "task_type": "lexical_detection", "query": "rate", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"rate\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.64, 0.95, "rate"]], "response_mode": "offline", "answer": "[{\"word\":\"rate\",\"start\":0.64,\"end\":0.95}]"} {"utterance_id": "3575-170457-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0051.flac", "task_type": "semantic_counting", "query": "food", "prompt": "When the utterance is complete, report how a food name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 0.62, "one"], [2.54, 2.8, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.45,\"end\":0.62},{\"word\":\"two\",\"start\":2.54,\"end\":2.8}]"} {"utterance_id": "1320-122617-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0033.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.39, 5.8, "uncas"]], "response_mode": "offline", "answer": "[{\"word\":\"uncas\",\"start\":5.39,\"end\":5.8}]"} {"utterance_id": "237-134493-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0003.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.05, 3.58, "dozen"]], "response_mode": "offline", "answer": "[{\"word\":\"dozen\",\"start\":3.05,\"end\":3.58}]"} {"utterance_id": "1580-141084-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0032.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.21, 2.55, "hand"], [4.04, 4.67, "features"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":2.21,\"end\":2.55},{\"word\":\"features\",\"start\":4.04,\"end\":4.67}]"} {"utterance_id": "121-127105-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0036.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.56, 2.81, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.56,\"end\":2.81}]"} {"utterance_id": "4077-13751-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0011.flac", "task_type": "semantic_counting", "query": "state_or_province", "prompt": "Once the audio has finished, list each occurrence of a state or province name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[3.39, 3.93, "one"], [5.41, 6.22, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.39,\"end\":3.93},{\"word\":\"two\",\"start\":5.41,\"end\":6.22}]"} {"utterance_id": "5683-32865-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0012.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.46, 0.89, "wylder"], [10.42, 10.82, "lake"], [18.68, 19.55, "missus dorothy"]], "response_mode": "offline", "answer": "[{\"word\":\"wylder\",\"start\":0.46,\"end\":0.89},{\"word\":\"lake\",\"start\":10.42,\"end\":10.82},{\"word\":\"missus dorothy\",\"start\":18.68,\"end\":19.55}]"} {"utterance_id": "1221-135767-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0008.flac", "task_type": "lexical_detection", "query": "therefore", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"therefore\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.6, 1.11, "therefore"]], "response_mode": "offline", "answer": "[{\"word\":\"therefore\",\"start\":0.6,\"end\":1.11}]"} {"utterance_id": "3575-170457-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0035.flac", "task_type": "lexical_detection", "query": "wear", "prompt": "Process the full audio offline. Find every instance of the word \"wear\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.27, 8.52, "wear"]], "response_mode": "offline", "answer": "[{\"word\":\"wear\",\"start\":8.27,\"end\":8.52}]"} {"utterance_id": "7729-102255-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0042.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.31, 2.76, "governor"]], "response_mode": "offline", "answer": "[{\"word\":\"governor\",\"start\":2.31,\"end\":2.76}]"} {"utterance_id": "6829-68771-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0031.flac", "task_type": "lexical_detection", "query": "hands", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"hands\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.81, 2.23, "hands"]], "response_mode": "offline", "answer": "[{\"word\":\"hands\",\"start\":1.81,\"end\":2.23}]"} {"utterance_id": "6930-76324-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0001.flac", "task_type": "lexical_detection", "query": "nearer", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"nearer\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.31, 1.62, "nearer"]], "response_mode": "offline", "answer": "[{\"word\":\"nearer\",\"start\":1.31,\"end\":1.62}]"} {"utterance_id": "2961-961-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0003.flac", "task_type": "lexical_detection", "query": "approve", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"approve\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.7, 4.25, "approve"]], "response_mode": "offline", "answer": "[{\"word\":\"approve\",\"start\":3.7,\"end\":4.25}]"} {"utterance_id": "2961-960-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0006.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen carefully to the entire clip, then report every instance of an ordinal number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.76, 10.25, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":9.76,\"end\":10.25}]"} {"utterance_id": "2094-142345-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0047.flac", "task_type": "lexical_detection", "query": "home", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"home\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.29, 1.62, "home"]], "response_mode": "offline", "answer": "[{\"word\":\"home\",\"start\":1.29,\"end\":1.62}]"} {"utterance_id": "7021-85628-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0001.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.13, 3.62, "back"]], "response_mode": "offline", "answer": "[{\"word\":\"back\",\"start\":3.13,\"end\":3.62}]"} {"utterance_id": "3729-6852-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0004.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Once the audio has finished, list each occurrence of a body part as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[0.75, 1.34, "one"], [11.22, 11.84, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.75,\"end\":1.34},{\"word\":\"two\",\"start\":11.22,\"end\":11.84}]"} {"utterance_id": "121-127105-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0009.flac", "task_type": "lexical_detection", "query": "dead", "prompt": "After the audio ends, extract all occurrences of the word \"dead\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.69, 0.96, "dead"]], "response_mode": "offline", "answer": "[{\"word\":\"dead\",\"start\":0.69,\"end\":0.96}]"} {"utterance_id": "6829-68771-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0019.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Do not answer until the audio ends. Then list each instance of a city name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.38, 3.99, "elmhurst"]], "response_mode": "offline", "answer": "[{\"word\":\"elmhurst\",\"start\":3.38,\"end\":3.99}]"} {"utterance_id": "2830-3980-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0028.flac", "task_type": "lexical_detection", "query": "shutting", "prompt": "When the utterance is complete, report all instances of the word \"shutting\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.4, 1.75, "shutting"]], "response_mode": "offline", "answer": "[{\"word\":\"shutting\",\"start\":1.4,\"end\":1.75}]"} {"utterance_id": "4507-16021-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0044.flac", "task_type": "lexical_detection", "query": "closely", "prompt": "After listening to the whole recording, output every instance of the word \"closely\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 1.25, "closely"]], "response_mode": "offline", "answer": "[{\"word\":\"closely\",\"start\":0.53,\"end\":1.25}]"} {"utterance_id": "4507-16021-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0009.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a disease name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.18, 2.74, "wound"]], "response_mode": "offline", "answer": "[{\"word\":\"wound\",\"start\":2.18,\"end\":2.74}]"} {"utterance_id": "3575-170457-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0005.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.06, 4.34, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.06,\"end\":4.34}]"} {"utterance_id": "7176-92135-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0012.flac", "task_type": "lexical_detection", "query": "clever", "prompt": "Once the audio has finished, identify every instance of the word \"clever\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.01, 4.39, "clever"]], "response_mode": "offline", "answer": "[{\"word\":\"clever\",\"start\":4.01,\"end\":4.39}]"} {"utterance_id": "4507-16021-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0042.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[6.48, 6.72, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":6.48,\"end\":6.72}]"} {"utterance_id": "4077-13754-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0004.flac", "task_type": "lexical_detection", "query": "west", "prompt": "Process the full audio offline. Find every instance of the word \"west\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.21, 4.65, "west"]], "response_mode": "offline", "answer": "[{\"word\":\"west\",\"start\":4.21,\"end\":4.65}]"} {"utterance_id": "121-121726-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0001.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.14, 5.66, "tongue"]], "response_mode": "offline", "answer": "[{\"word\":\"tongue\",\"start\":5.14,\"end\":5.66}]"} {"utterance_id": "6829-68769-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0000.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a money or currency term with fields word, start, end (seconds). If none, reply [].", "labels": [[11.14, 11.51, "check"]], "response_mode": "offline", "answer": "[{\"word\":\"check\",\"start\":11.14,\"end\":11.51}]"} {"utterance_id": "7729-102255-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0039.flac", "task_type": "lexical_detection", "query": "balls", "prompt": "Once the audio has finished, identify every instance of the word \"balls\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.59, 4.95, "balls"]], "response_mode": "offline", "answer": "[{\"word\":\"balls\",\"start\":4.59,\"end\":4.95}]"} {"utterance_id": "5142-36377-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0001.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.27, 0.62, "five"]], "response_mode": "offline", "answer": "[{\"word\":\"five\",\"start\":0.27,\"end\":0.62}]"} {"utterance_id": "3575-170457-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0053.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.65, 0.9, "anne"], [3.2, 3.6, "charlotte"]], "response_mode": "offline", "answer": "[{\"word\":\"anne\",\"start\":0.65,\"end\":0.9},{\"word\":\"charlotte\",\"start\":3.2,\"end\":3.6}]"} {"utterance_id": "1320-122617-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0032.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.25, 5.39, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.25,\"end\":5.39}]"} {"utterance_id": "5142-36377-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.27, 2.68, "dumas"]], "response_mode": "offline", "answer": "[{\"word\":\"dumas\",\"start\":2.27,\"end\":2.68}]"} {"utterance_id": "3570-5694-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0022.flac", "task_type": "lexical_detection", "query": "required", "prompt": "After the audio ends, extract all occurrences of the word \"required\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.94, 3.38, "required"]], "response_mode": "offline", "answer": "[{\"word\":\"required\",\"start\":2.94,\"end\":3.38}]"} {"utterance_id": "2961-961-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0012.flac", "task_type": "lexical_detection", "query": "cities", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"cities\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.61, 6.21, "cities"]], "response_mode": "offline", "answer": "[{\"word\":\"cities\",\"start\":5.61,\"end\":6.21}]"} {"utterance_id": "61-70968-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0029.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.36, 0.81, "squire"]], "response_mode": "offline", "answer": "[{\"word\":\"squire\",\"start\":0.36,\"end\":0.81}]"} {"utterance_id": "4970-29095-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0027.flac", "task_type": "lexical_detection", "query": "sleek", "prompt": "Process the full audio offline. Find every instance of the word \"sleek\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.82, 5.19, "sleek"]], "response_mode": "offline", "answer": "[{\"word\":\"sleek\",\"start\":4.82,\"end\":5.19}]"} {"utterance_id": "3729-6852-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0029.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "When the utterance is complete, report all instances of an ordinal number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.6, 3.01, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":2.6,\"end\":3.01}]"} {"utterance_id": "2830-3979-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0010.flac", "task_type": "lexical_detection", "query": "luther's", "prompt": "After the audio ends, extract all occurrences of the word \"luther's\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.46, 1.86, "luther's"]], "response_mode": "offline", "answer": "[{\"word\":\"luther's\",\"start\":1.46,\"end\":1.86}]"} {"utterance_id": "1320-122617-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0038.flac", "task_type": "lexical_detection", "query": "groans", "prompt": "Once the audio has finished, identify every instance of the word \"groans\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.25, 1.9, "groans"]], "response_mode": "offline", "answer": "[{\"word\":\"groans\",\"start\":1.25,\"end\":1.9}]"} {"utterance_id": "1089-134686-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0011.flac", "task_type": "semantic_detection", "query": "date", "prompt": "When the utterance is complete, report all instances of a date including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.7, 1.14, "saturday"]], "response_mode": "offline", "answer": "[{\"word\":\"saturday\",\"start\":0.7,\"end\":1.14}]"} {"utterance_id": "8555-292519-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0008.flac", "task_type": "lexical_detection", "query": "city", "prompt": "After the audio ends, extract all occurrences of the word \"city\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.67, 2.04, "city"]], "response_mode": "offline", "answer": "[{\"word\":\"city\",\"start\":1.67,\"end\":2.04}]"} {"utterance_id": "2830-3980-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0007.flac", "task_type": "lexical_counting", "query": "office", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"office\" (word, start, end). If none, reply [].", "labels": [[2.62, 3.15, "one"], [7.14, 7.61, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.62,\"end\":3.15},{\"word\":\"two\",\"start\":7.14,\"end\":7.61}]"} {"utterance_id": "908-157963-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0002.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Once the audio has finished, identify every instance of a plant name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.1, 1.58, "lotus"]], "response_mode": "offline", "answer": "[{\"word\":\"lotus\",\"start\":1.1,\"end\":1.58}]"} {"utterance_id": "6829-68769-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0043.flac", "task_type": "lexical_detection", "query": "prison", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"prison\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.69, 2.11, "prison"]], "response_mode": "offline", "answer": "[{\"word\":\"prison\",\"start\":1.69,\"end\":2.11}]"} {"utterance_id": "237-126133-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0005.flac", "task_type": "lexical_detection", "query": "van", "prompt": "Once the audio has finished, identify every instance of the word \"van\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.2, 2.57, "van"]], "response_mode": "offline", "answer": "[{\"word\":\"van\",\"start\":2.2,\"end\":2.57}]"} {"utterance_id": "8230-279154-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0040.flac", "task_type": "lexical_detection", "query": "inadequate", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"inadequate\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.54, 6.39, "inadequate"]], "response_mode": "offline", "answer": "[{\"word\":\"inadequate\",\"start\":5.54,\"end\":6.39}]"} {"utterance_id": "6829-68769-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0025.flac", "task_type": "lexical_detection", "query": "rogers", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"rogers\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.49, 0.92, "rogers"]], "response_mode": "offline", "answer": "[{\"word\":\"rogers\",\"start\":0.49,\"end\":0.92}]"} {"utterance_id": "7127-75946-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0017.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "After listening to the whole recording, output timed count words for a body part (one, two, three, ...) as JSON. If none, reply [].", "labels": [[0.53, 1.16, "one"], [5.36, 5.68, "two"], [11.14, 11.45, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.53,\"end\":1.16},{\"word\":\"two\",\"start\":5.36,\"end\":5.68},{\"word\":\"three\",\"start\":11.14,\"end\":11.45}]"} {"utterance_id": "8463-294825-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0005.flac", "task_type": "lexical_detection", "query": "inside", "prompt": "After the audio ends, extract all occurrences of the word \"inside\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.66, 2.23, "inside"]], "response_mode": "offline", "answer": "[{\"word\":\"inside\",\"start\":1.66,\"end\":2.23}]"} {"utterance_id": "61-70968-0056", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0056.flac", "task_type": "lexical_detection", "query": "wine", "prompt": "Once the audio has finished, identify every instance of the word \"wine\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.52, 0.75, "wine"]], "response_mode": "offline", "answer": "[{\"word\":\"wine\",\"start\":0.52,\"end\":0.75}]"} {"utterance_id": "4992-41806-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0011.flac", "task_type": "lexical_detection", "query": "carey", "prompt": "After the audio ends, extract all occurrences of the word \"carey\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.71, 1.1, "carey"]], "response_mode": "offline", "answer": "[{\"word\":\"carey\",\"start\":0.71,\"end\":1.1}]"} {"utterance_id": "1580-141083-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0006.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.85, 4.31, "servant"]], "response_mode": "offline", "answer": "[{\"word\":\"servant\",\"start\":3.85,\"end\":4.31}]"} {"utterance_id": "1580-141084-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0021.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.3, 0.74, "palm"]], "response_mode": "offline", "answer": "[{\"word\":\"palm\",\"start\":0.3,\"end\":0.74}]"} {"utterance_id": "2300-131720-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0038.flac", "task_type": "lexical_detection", "query": "suit", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"suit\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.36, 4.61, "suit"]], "response_mode": "offline", "answer": "[{\"word\":\"suit\",\"start\":4.36,\"end\":4.61}]"} {"utterance_id": "237-126133-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0010.flac", "task_type": "lexical_detection", "query": "dearest", "prompt": "After listening to the whole recording, output every instance of the word \"dearest\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.26, 1.75, "dearest"]], "response_mode": "offline", "answer": "[{\"word\":\"dearest\",\"start\":1.26,\"end\":1.75}]"} {"utterance_id": "8455-210777-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0023.flac", "task_type": "lexical_detection", "query": "officers", "prompt": "Once the audio has finished, identify every instance of the word \"officers\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.06, 1.61, "officers"]], "response_mode": "offline", "answer": "[{\"word\":\"officers\",\"start\":1.06,\"end\":1.61}]"} {"utterance_id": "237-134500-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0025.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.8, 1.39, "emil"]], "response_mode": "offline", "answer": "[{\"word\":\"emil\",\"start\":0.8,\"end\":1.39}]"} {"utterance_id": "121-127105-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0029.flac", "task_type": "lexical_detection", "query": "governess", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"governess\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.62, 5.17, "governess"]], "response_mode": "offline", "answer": "[{\"word\":\"governess\",\"start\":4.62,\"end\":5.17}]"} {"utterance_id": "7176-88083-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0026.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Listen to the entire clip, then report the count sequence for a body part as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.21, 1.58, "one"], [2.85, 3.18, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.21,\"end\":1.58},{\"word\":\"two\",\"start\":2.85,\"end\":3.18}]"} {"utterance_id": "2830-3980-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.22, 0.76, "paul"]], "response_mode": "offline", "answer": "[{\"word\":\"paul\",\"start\":0.22,\"end\":0.76}]"} {"utterance_id": "1089-134686-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0010.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.52, 2.93, "head"]], "response_mode": "offline", "answer": "[{\"word\":\"head\",\"start\":2.52,\"end\":2.93}]"} {"utterance_id": "4992-41797-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0007.flac", "task_type": "lexical_detection", "query": "books", "prompt": "After the audio ends, extract all occurrences of the word \"books\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.92, 6.47, "books"]], "response_mode": "offline", "answer": "[{\"word\":\"books\",\"start\":5.92,\"end\":6.47}]"} {"utterance_id": "237-134493-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0016.flac", "task_type": "lexical_detection", "query": "road", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"road\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.31, 1.78, "road"]], "response_mode": "offline", "answer": "[{\"word\":\"road\",\"start\":1.31,\"end\":1.78}]"} {"utterance_id": "1188-133604-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0031.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.59, 0.89, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.59,\"end\":0.89}]"} {"utterance_id": "7729-102255-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0021.flac", "task_type": "lexical_detection", "query": "affair", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"affair\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.55, 1.12, "affair"]], "response_mode": "offline", "answer": "[{\"word\":\"affair\",\"start\":0.55,\"end\":1.12}]"} {"utterance_id": "7176-92135-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0022.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.88, 7.33, "hamlet"]], "response_mode": "offline", "answer": "[{\"word\":\"hamlet\",\"start\":6.88,\"end\":7.33}]"} {"utterance_id": "908-31957-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0015.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Listen to the full audio. Count every occurrence of a body part. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.22, 3.84, "one"], [8.52, 9.0, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.22,\"end\":3.84},{\"word\":\"two\",\"start\":8.52,\"end\":9.0}]"} {"utterance_id": "1221-135767-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0013.flac", "task_type": "lexical_detection", "query": "lifting", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"lifting\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.39, 0.78, "lifting"]], "response_mode": "offline", "answer": "[{\"word\":\"lifting\",\"start\":0.39,\"end\":0.78}]"} {"utterance_id": "7176-92135-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0018.flac", "task_type": "lexical_detection", "query": "modern", "prompt": "After the audio ends, extract all occurrences of the word \"modern\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.47, 0.88, "modern"]], "response_mode": "offline", "answer": "[{\"word\":\"modern\",\"start\":0.47,\"end\":0.88}]"} {"utterance_id": "5105-28233-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28233/5105-28233-0004.flac", "task_type": "lexical_detection", "query": "infantry", "prompt": "After the audio ends, extract all occurrences of the word \"infantry\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.9, 3.48, "infantry"]], "response_mode": "offline", "answer": "[{\"word\":\"infantry\",\"start\":2.9,\"end\":3.48}]"} {"utterance_id": "3729-6852-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0010.flac", "task_type": "lexical_detection", "query": "family", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"family\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.71, 2.37, "family"]], "response_mode": "offline", "answer": "[{\"word\":\"family\",\"start\":1.71,\"end\":2.37}]"} {"utterance_id": "5142-33396-0049", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0049.flac", "task_type": "lexical_detection", "query": "friend", "prompt": "After the audio ends, extract all occurrences of the word \"friend\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.44, 0.86, "friend"]], "response_mode": "offline", "answer": "[{\"word\":\"friend\",\"start\":0.44,\"end\":0.86}]"} {"utterance_id": "237-134500-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0019.flac", "task_type": "lexical_detection", "query": "lap", "prompt": "After listening to the whole recording, output every instance of the word \"lap\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.93, 2.31, "lap"]], "response_mode": "offline", "answer": "[{\"word\":\"lap\",\"start\":1.93,\"end\":2.31}]"} {"utterance_id": "260-123286-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0025.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Do not answer until the audio ends. Then list each instance of an animal name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.69, 3.33, "reptiles"]], "response_mode": "offline", "answer": "[{\"word\":\"reptiles\",\"start\":2.69,\"end\":3.33}]"} {"utterance_id": "237-126133-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0014.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.64, 1.32, "phronsie"], [3.51, 3.94, "polly's"], [5.61, 6.09, "jasper's"]], "response_mode": "offline", "answer": "[{\"word\":\"phronsie\",\"start\":0.64,\"end\":1.32},{\"word\":\"polly's\",\"start\":3.51,\"end\":3.94},{\"word\":\"jasper's\",\"start\":5.61,\"end\":6.09}]"} {"utterance_id": "6930-75918-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0015.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a country name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.91, 4.47, "country's"]], "response_mode": "offline", "answer": "[{\"word\":\"country's\",\"start\":3.91,\"end\":4.47}]"} {"utterance_id": "4970-29093-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0016.flac", "task_type": "lexical_detection", "query": "soon", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"soon\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.42, 1.88, "soon"]], "response_mode": "offline", "answer": "[{\"word\":\"soon\",\"start\":1.42,\"end\":1.88}]"} {"utterance_id": "1089-134686-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0018.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen carefully to the entire clip, then report every instance of an organization name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.71, 8.13, "church"]], "response_mode": "offline", "answer": "[{\"word\":\"church\",\"start\":7.71,\"end\":8.13}]"} {"utterance_id": "3575-170457-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0046.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "After the audio ends, extract all occurrences of a disease name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[17.92, 18.31, "sickness"]], "response_mode": "offline", "answer": "[{\"word\":\"sickness\",\"start\":17.92,\"end\":18.31}]"} {"utterance_id": "7729-102255-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0037.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "When the utterance is complete, report all instances of a time of day including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.73, 6.56, "five o'clock"]], "response_mode": "offline", "answer": "[{\"word\":\"five o'clock\",\"start\":5.73,\"end\":6.56}]"} {"utterance_id": "3575-170457-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0042.flac", "task_type": "lexical_detection", "query": "doubtful", "prompt": "Process the full audio offline. Find every instance of the word \"doubtful\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.0, 10.54, "doubtful"]], "response_mode": "offline", "answer": "[{\"word\":\"doubtful\",\"start\":10.0,\"end\":10.54}]"} {"utterance_id": "4446-2273-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0010.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[5.57, 5.83, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":5.57,\"end\":5.83}]"} {"utterance_id": "7176-92135-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0023.flac", "task_type": "lexical_counting", "query": "double", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"double\" (word, start, end). If none, reply [].", "labels": [[0.55, 0.92, "one"], [1.95, 2.2, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.55,\"end\":0.92},{\"word\":\"two\",\"start\":1.95,\"end\":2.2}]"} {"utterance_id": "260-123286-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0017.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Process the full audio offline. Find every instance of an animal name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.03, 2.56, "monsters"]], "response_mode": "offline", "answer": "[{\"word\":\"monsters\",\"start\":2.03,\"end\":2.56}]"} {"utterance_id": "6829-68769-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0002.flac", "task_type": "lexical_detection", "query": "old", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"old\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.31, 2.56, "old"]], "response_mode": "offline", "answer": "[{\"word\":\"old\",\"start\":2.31,\"end\":2.56}]"} {"utterance_id": "260-123286-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0009.flac", "task_type": "lexical_detection", "query": "lips", "prompt": "After listening to the whole recording, output every instance of the word \"lips\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.05, 4.48, "lips"]], "response_mode": "offline", "answer": "[{\"word\":\"lips\",\"start\":4.05,\"end\":4.48}]"} {"utterance_id": "6829-68769-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0039.flac", "task_type": "lexical_detection", "query": "seated", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"seated\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.13, 3.6, "seated"]], "response_mode": "offline", "answer": "[{\"word\":\"seated\",\"start\":3.13,\"end\":3.6}]"} {"utterance_id": "1995-1826-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0000.flac", "task_type": "lexical_detection", "query": "fifteenth", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"fifteenth\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.09, 3.53, "fifteenth"]], "response_mode": "offline", "answer": "[{\"word\":\"fifteenth\",\"start\":3.09,\"end\":3.53}]"} {"utterance_id": "908-157963-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0025.flac", "task_type": "lexical_detection", "query": "mothers", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"mothers\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[7.28, 7.83, "mothers"]], "response_mode": "offline", "answer": "[{\"word\":\"mothers\",\"start\":7.28,\"end\":7.83}]"} {"utterance_id": "5142-33396-0064", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0064.flac", "task_type": "semantic_counting", "query": "kinship", "prompt": "After the audio ends, tally every instance of a family relation. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.48, 0.85, "one"], [2.06, 2.32, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.48,\"end\":0.85},{\"word\":\"two\",\"start\":2.06,\"end\":2.32}]"} {"utterance_id": "672-122797-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0024.flac", "task_type": "lexical_detection", "query": "departure", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"departure\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.56, 1.32, "departure"]], "response_mode": "offline", "answer": "[{\"word\":\"departure\",\"start\":0.56,\"end\":1.32}]"} {"utterance_id": "5142-33396-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0027.flac", "task_type": "lexical_detection", "query": "acts", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"acts\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.27, 0.49, "acts"]], "response_mode": "offline", "answer": "[{\"word\":\"acts\",\"start\":0.27,\"end\":0.49}]"} {"utterance_id": "6930-81414-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0027.flac", "task_type": "lexical_detection", "query": "remembered", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"remembered\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.92, 2.51, "remembered"]], "response_mode": "offline", "answer": "[{\"word\":\"remembered\",\"start\":1.92,\"end\":2.51}]"} {"utterance_id": "8224-274381-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0014.flac", "task_type": "lexical_detection", "query": "unwilling", "prompt": "After the audio ends, extract all occurrences of the word \"unwilling\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.23, 5.9, "unwilling"]], "response_mode": "offline", "answer": "[{\"word\":\"unwilling\",\"start\":5.23,\"end\":5.9}]"} {"utterance_id": "7176-88083-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0013.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[5.29, 5.63, "heart"], [12.61, 12.7, "eye"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":5.29,\"end\":5.63},{\"word\":\"eye\",\"start\":12.61,\"end\":12.7}]"} {"utterance_id": "6829-68771-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0013.flac", "task_type": "lexical_detection", "query": "fete", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"fete\" with fields word, start, end (seconds). If none, reply [].", "labels": [[6.41, 6.76, "fete"]], "response_mode": "offline", "answer": "[{\"word\":\"fete\",\"start\":6.41,\"end\":6.76}]"} {"utterance_id": "8230-279154-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0027.flac", "task_type": "lexical_detection", "query": "suitable", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"suitable\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.92, 2.38, "suitable"]], "response_mode": "offline", "answer": "[{\"word\":\"suitable\",\"start\":1.92,\"end\":2.38}]"} {"utterance_id": "5639-40744-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0016.flac", "task_type": "lexical_detection", "query": "resolved", "prompt": "After the audio ends, extract all occurrences of the word \"resolved\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.22, 1.71, "resolved"]], "response_mode": "offline", "answer": "[{\"word\":\"resolved\",\"start\":1.22,\"end\":1.71}]"} {"utterance_id": "61-70968-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0032.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.15, 1.52, "wizard's"], [2.76, 3.19, "squire"]], "response_mode": "offline", "answer": "[{\"word\":\"wizard's\",\"start\":1.15,\"end\":1.52},{\"word\":\"squire\",\"start\":2.76,\"end\":3.19}]"} {"utterance_id": "7021-85628-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0014.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.64, 2.05, "head"]], "response_mode": "offline", "answer": "[{\"word\":\"head\",\"start\":1.64,\"end\":2.05}]"} {"utterance_id": "6930-75918-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0009.flac", "task_type": "lexical_detection", "query": "raoul", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"raoul\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.76, 2.16, "raoul"]], "response_mode": "offline", "answer": "[{\"word\":\"raoul\",\"start\":1.76,\"end\":2.16}]"} {"utterance_id": "61-70970-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0037.flac", "task_type": "lexical_detection", "query": "pleasantly", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"pleasantly\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.25, 1.77, "pleasantly"]], "response_mode": "offline", "answer": "[{\"word\":\"pleasantly\",\"start\":1.25,\"end\":1.77}]"} {"utterance_id": "1188-133604-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0011.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After the audio ends, extract all occurrences of a color name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.04, 14.24, "gray"]], "response_mode": "offline", "answer": "[{\"word\":\"gray\",\"start\":14.04,\"end\":14.24}]"} {"utterance_id": "8455-210777-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0022.flac", "task_type": "lexical_detection", "query": "beasts", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"beasts\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.47, 4.07, "beasts"]], "response_mode": "offline", "answer": "[{\"word\":\"beasts\",\"start\":3.47,\"end\":4.07}]"} {"utterance_id": "61-70968-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0013.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.31, 2.77, "sons"]], "response_mode": "offline", "answer": "[{\"word\":\"sons\",\"start\":2.31,\"end\":2.77}]"} {"utterance_id": "260-123286-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0007.flac", "task_type": "lexical_detection", "query": "voyage", "prompt": "Process the full audio offline. Find every instance of the word \"voyage\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.6, 3.03, "voyage"]], "response_mode": "offline", "answer": "[{\"word\":\"voyage\",\"start\":2.6,\"end\":3.03}]"} {"utterance_id": "672-122797-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0027.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.59, 1.25, "servants"]], "response_mode": "offline", "answer": "[{\"word\":\"servants\",\"start\":0.59,\"end\":1.25}]"} {"utterance_id": "3575-170457-0055", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0055.flac", "task_type": "lexical_detection", "query": "perception", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"perception\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.59, 3.18, "perception"]], "response_mode": "offline", "answer": "[{\"word\":\"perception\",\"start\":2.59,\"end\":3.18}]"} {"utterance_id": "1188-133604-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0019.flac", "task_type": "lexical_detection", "query": "feeble", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"feeble\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.16, 3.62, "feeble"]], "response_mode": "offline", "answer": "[{\"word\":\"feeble\",\"start\":3.16,\"end\":3.62}]"} {"utterance_id": "2300-131720-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0018.flac", "task_type": "lexical_detection", "query": "commercial", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"commercial\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.77, 5.28, "commercial"]], "response_mode": "offline", "answer": "[{\"word\":\"commercial\",\"start\":4.77,\"end\":5.28}]"} {"utterance_id": "61-70970-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0022.flac", "task_type": "lexical_detection", "query": "hut", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"hut\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.2, 1.46, "hut"]], "response_mode": "offline", "answer": "[{\"word\":\"hut\",\"start\":1.2,\"end\":1.46}]"} {"utterance_id": "4077-13754-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0013.flac", "task_type": "lexical_detection", "query": "administration", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"administration\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.93, 2.72, "administration"]], "response_mode": "offline", "answer": "[{\"word\":\"administration\",\"start\":1.93,\"end\":2.72}]"} {"utterance_id": "4507-16021-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0042.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Do not answer until the audio ends. Then list each instance of a color name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.77, 1.41, "black"], [3.34, 3.84, "blacker"], [6.72, 7.85, "blacknesses"]], "response_mode": "offline", "answer": "[{\"word\":\"black\",\"start\":0.77,\"end\":1.41},{\"word\":\"blacker\",\"start\":3.34,\"end\":3.84},{\"word\":\"blacknesses\",\"start\":6.72,\"end\":7.85}]"} {"utterance_id": "6829-68771-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0015.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.65, 4.08, "maids"]], "response_mode": "offline", "answer": "[{\"word\":\"maids\",\"start\":3.65,\"end\":4.08}]"} {"utterance_id": "1284-1181-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0003.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a color name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.02, 1.38, "brown"]], "response_mode": "offline", "answer": "[{\"word\":\"brown\",\"start\":1.02,\"end\":1.38}]"} {"utterance_id": "1995-1826-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0018.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a color name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.52, 1.85, "green"]], "response_mode": "offline", "answer": "[{\"word\":\"green\",\"start\":1.52,\"end\":1.85}]"} {"utterance_id": "3729-6852-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0018.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.23, 3.56, "waiter"]], "response_mode": "offline", "answer": "[{\"word\":\"waiter\",\"start\":3.23,\"end\":3.56}]"} {"utterance_id": "2830-3980-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0013.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.53, 1.89, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":1.53,\"end\":1.89}]"} {"utterance_id": "672-122797-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0035.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.66, 2.08, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.66,\"end\":2.08}]"} {"utterance_id": "7021-79759-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79759/7021-79759-0001.flac", "task_type": "lexical_detection", "query": "comparatively", "prompt": "When the utterance is complete, report all instances of the word \"comparatively\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.08, 1.86, "comparatively"]], "response_mode": "offline", "answer": "[{\"word\":\"comparatively\",\"start\":1.08,\"end\":1.86}]"} {"utterance_id": "3570-5695-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0002.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Once the audio has finished, identify every instance of a family relation. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.83, 7.15, "wife"]], "response_mode": "offline", "answer": "[{\"word\":\"wife\",\"start\":6.83,\"end\":7.15}]"} {"utterance_id": "1221-135767-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0002.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a family relation with fields word, start, end (seconds). If none, reply [].", "labels": [[9.72, 10.21, "child"]], "response_mode": "offline", "answer": "[{\"word\":\"child\",\"start\":9.72,\"end\":10.21}]"} {"utterance_id": "61-70970-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0033.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen carefully to the entire clip, then report every instance of a city name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.34, 2.81, "nottingham"]], "response_mode": "offline", "answer": "[{\"word\":\"nottingham\",\"start\":2.34,\"end\":2.81}]"} {"utterance_id": "61-70968-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0010.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[7.12, 7.36, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.12,\"end\":7.36}]"} {"utterance_id": "3570-5694-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0013.flac", "task_type": "lexical_counting", "query": "inheritance", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"inheritance\" with word/start/end. If none, reply [].", "labels": [[2.05, 2.63, "one"], [3.96, 4.49, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.05,\"end\":2.63},{\"word\":\"two\",\"start\":3.96,\"end\":4.49}]"} {"utterance_id": "237-126133-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0014.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "When the utterance is complete, report how a body part accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.94, 4.4, "one"], [6.09, 6.58, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.94,\"end\":4.4},{\"word\":\"two\",\"start\":6.09,\"end\":6.58}]"} {"utterance_id": "7127-75947-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0003.flac", "task_type": "lexical_detection", "query": "harmony", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"harmony\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.34, 4.76, "harmony"]], "response_mode": "offline", "answer": "[{\"word\":\"harmony\",\"start\":4.34,\"end\":4.76}]"} {"utterance_id": "672-122797-0056", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0056.flac", "task_type": "lexical_detection", "query": "thinking", "prompt": "After listening to the whole recording, output every instance of the word \"thinking\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.75, 2.2, "thinking"]], "response_mode": "offline", "answer": "[{\"word\":\"thinking\",\"start\":1.75,\"end\":2.2}]"} {"utterance_id": "1580-141083-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0048.flac", "task_type": "lexical_detection", "query": "key", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"key\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.66, 1.9, "key"]], "response_mode": "offline", "answer": "[{\"word\":\"key\",\"start\":1.66,\"end\":1.9}]"} {"utterance_id": "1580-141084-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0046.flac", "task_type": "lexical_detection", "query": "truth", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"truth\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.93, 1.27, "truth"]], "response_mode": "offline", "answer": "[{\"word\":\"truth\",\"start\":0.93,\"end\":1.27}]"} {"utterance_id": "8555-284447-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0009.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Process the full audio offline. Find every instance of a time of day and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.24, 0.73, "mornin"]], "response_mode": "offline", "answer": "[{\"word\":\"mornin\",\"start\":0.24,\"end\":0.73}]"} {"utterance_id": "1580-141084-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0031.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.06, 1.92, "mister gilchrist"]], "response_mode": "offline", "answer": "[{\"word\":\"mister gilchrist\",\"start\":1.06,\"end\":1.92}]"} {"utterance_id": "1089-134686-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0016.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.48, 3.02, "dedalus"]], "response_mode": "offline", "answer": "[{\"word\":\"dedalus\",\"start\":2.48,\"end\":3.02}]"} {"utterance_id": "6829-68771-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0028.flac", "task_type": "lexical_detection", "query": "attracted", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"attracted\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.79, 3.39, "attracted"]], "response_mode": "offline", "answer": "[{\"word\":\"attracted\",\"start\":2.79,\"end\":3.39}]"} {"utterance_id": "8555-284449-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0017.flac", "task_type": "semantic_detection", "query": "country", "prompt": "You will hear a complete utterance. When it ends, list every instance of a country name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.4, 10.74, "country"]], "response_mode": "offline", "answer": "[{\"word\":\"country\",\"start\":10.4,\"end\":10.74}]"} {"utterance_id": "672-122797-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0050.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a plant name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.6, 2.39, "fir tree"], [3.68, 4.44, "branches"]], "response_mode": "offline", "answer": "[{\"word\":\"fir tree\",\"start\":1.6,\"end\":2.39},{\"word\":\"branches\",\"start\":3.68,\"end\":4.44}]"} {"utterance_id": "260-123440-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0020.flac", "task_type": "lexical_detection", "query": "won't", "prompt": "After the audio ends, extract all occurrences of the word \"won't\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.35, 0.57, "won't"]], "response_mode": "offline", "answer": "[{\"word\":\"won't\",\"start\":0.35,\"end\":0.57}]"} {"utterance_id": "1188-133604-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0003.flac", "task_type": "lexical_detection", "query": "student", "prompt": "Once the audio has finished, identify every instance of the word \"student\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.37, 6.82, "student"]], "response_mode": "offline", "answer": "[{\"word\":\"student\",\"start\":6.37,\"end\":6.82}]"} {"utterance_id": "4077-13751-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0007.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Process the full audio offline. Find every instance of a color name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.03, 7.36, "dark"]], "response_mode": "offline", "answer": "[{\"word\":\"dark\",\"start\":7.03,\"end\":7.36}]"} {"utterance_id": "8224-274381-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0004.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.34, 1.08, "five hundred"], [15.25, 16.16, "six thousand"]], "response_mode": "offline", "answer": "[{\"word\":\"five hundred\",\"start\":0.34,\"end\":1.08},{\"word\":\"six thousand\",\"start\":15.25,\"end\":16.16}]"} {"utterance_id": "7021-85628-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0020.flac", "task_type": "lexical_detection", "query": "arrow", "prompt": "Process the full audio offline. Find every instance of the word \"arrow\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.37, 1.76, "arrow"]], "response_mode": "offline", "answer": "[{\"word\":\"arrow\",\"start\":1.37,\"end\":1.76}]"} {"utterance_id": "5639-40744-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0030.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a person's name with word/start/end. If none, reply [].", "labels": [[0.96, 1.56, "one"], [13.09, 13.65, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.96,\"end\":1.56},{\"word\":\"two\",\"start\":13.09,\"end\":13.65}]"} {"utterance_id": "5142-36377-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0005.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.52, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":2.35,\"end\":2.52}]"} {"utterance_id": "237-126133-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0006.flac", "task_type": "lexical_detection", "query": "opened", "prompt": "Once the audio has finished, identify every instance of the word \"opened\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.93, 2.28, "opened"]], "response_mode": "offline", "answer": "[{\"word\":\"opened\",\"start\":1.93,\"end\":2.28}]"} {"utterance_id": "237-134493-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0018.flac", "task_type": "semantic_detection", "query": "color", "prompt": "When the utterance is complete, report all instances of a color name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.19, 1.57, "white"]], "response_mode": "offline", "answer": "[{\"word\":\"white\",\"start\":1.19,\"end\":1.57}]"} {"utterance_id": "3729-6852-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.57, 4.57, "abbe conti"]], "response_mode": "offline", "answer": "[{\"word\":\"abbe conti\",\"start\":3.57,\"end\":4.57}]"} {"utterance_id": "7127-75946-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.27, 3.77, "colbert"]], "response_mode": "offline", "answer": "[{\"word\":\"colbert\",\"start\":3.27,\"end\":3.77}]"} {"utterance_id": "237-134493-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0005.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.37, 5.09, "pine tree"]], "response_mode": "offline", "answer": "[{\"word\":\"pine tree\",\"start\":4.37,\"end\":5.09}]"} {"utterance_id": "4446-2275-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0045.flac", "task_type": "lexical_detection", "query": "we've", "prompt": "Once the audio has finished, identify every instance of the word \"we've\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.29, 0.52, "we've"]], "response_mode": "offline", "answer": "[{\"word\":\"we've\",\"start\":0.29,\"end\":0.52}]"} {"utterance_id": "61-70968-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0040.flac", "task_type": "lexical_detection", "query": "willingly", "prompt": "After the audio ends, extract all occurrences of the word \"willingly\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.73, 1.26, "willingly"]], "response_mode": "offline", "answer": "[{\"word\":\"willingly\",\"start\":0.73,\"end\":1.26}]"} {"utterance_id": "1221-135766-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0000.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.72, 12.18, "child"]], "response_mode": "offline", "answer": "[{\"word\":\"child\",\"start\":11.72,\"end\":12.18}]"} {"utterance_id": "8224-274381-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0013.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.67, 6.2, "officers"]], "response_mode": "offline", "answer": "[{\"word\":\"officers\",\"start\":5.67,\"end\":6.2}]"} {"utterance_id": "2094-142345-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0015.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.68, 2.08, "molly"]], "response_mode": "offline", "answer": "[{\"word\":\"molly\",\"start\":1.68,\"end\":2.08}]"} {"utterance_id": "5639-40744-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0003.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Do not answer until the audio ends. Then list each instance of an ordinal number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.22, 4.58, "next"]], "response_mode": "offline", "answer": "[{\"word\":\"next\",\"start\":4.22,\"end\":4.58}]"} {"utterance_id": "8555-284449-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0008.flac", "task_type": "lexical_detection", "query": "magic", "prompt": "Once the audio has finished, identify every instance of the word \"magic\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.95, 2.33, "magic"]], "response_mode": "offline", "answer": "[{\"word\":\"magic\",\"start\":1.95,\"end\":2.33}]"} {"utterance_id": "7021-79740-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0014.flac", "task_type": "lexical_detection", "query": "effectual", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"effectual\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.29, 3.94, "effectual"]], "response_mode": "offline", "answer": "[{\"word\":\"effectual\",\"start\":3.29,\"end\":3.94}]"} {"utterance_id": "260-123288-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0017.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "When the utterance is complete, report all instances of an ordinal number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.25, 4.86, "final"]], "response_mode": "offline", "answer": "[{\"word\":\"final\",\"start\":4.25,\"end\":4.86}]"} {"utterance_id": "237-126133-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0012.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After the audio ends, extract all occurrences of a color name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.07, 3.53, "brown"]], "response_mode": "offline", "answer": "[{\"word\":\"brown\",\"start\":3.07,\"end\":3.53}]"} {"utterance_id": "3575-170457-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0046.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "You will hear a complete utterance. When it ends, list every instance of a state or province name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.39, 3.82, "yorkshire"]], "response_mode": "offline", "answer": "[{\"word\":\"yorkshire\",\"start\":3.39,\"end\":3.82}]"} {"utterance_id": "7176-88083-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0024.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.1, 6.42, "beak"]], "response_mode": "offline", "answer": "[{\"word\":\"beak\",\"start\":6.1,\"end\":6.42}]"} {"utterance_id": "3729-6852-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0033.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a money or currency term with fields word, start, end (seconds). If none, reply [].", "labels": [[4.02, 4.67, "fortune"]], "response_mode": "offline", "answer": "[{\"word\":\"fortune\",\"start\":4.02,\"end\":4.67}]"} {"utterance_id": "4446-2273-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0030.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.39, 0.92, "alexander"]], "response_mode": "offline", "answer": "[{\"word\":\"alexander\",\"start\":0.39,\"end\":0.92}]"} {"utterance_id": "4446-2273-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0003.flac", "task_type": "semantic_detection", "query": "date", "prompt": "Listen carefully to the entire clip, then report every instance of a date as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.46, 2.77, "sunday"]], "response_mode": "offline", "answer": "[{\"word\":\"sunday\",\"start\":2.46,\"end\":2.77}]"} {"utterance_id": "5142-33396-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0051.flac", "task_type": "lexical_detection", "query": "sif", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"sif\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.72, 2.21, "sif"]], "response_mode": "offline", "answer": "[{\"word\":\"sif\",\"start\":1.72,\"end\":2.21}]"} {"utterance_id": "4077-13754-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0006.flac", "task_type": "semantic_counting", "query": "organization", "prompt": "Listen to the entire clip, then report the count sequence for an organization name as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.73, 1.42, "one"], [4.05, 4.55, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.73,\"end\":1.42},{\"word\":\"two\",\"start\":4.05,\"end\":4.55}]"} {"utterance_id": "5142-33396-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0040.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.31, 1.73, "thralls"]], "response_mode": "offline", "answer": "[{\"word\":\"thralls\",\"start\":1.31,\"end\":1.73}]"} {"utterance_id": "8555-284449-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0016.flac", "task_type": "lexical_detection", "query": "dangerous", "prompt": "After the audio ends, extract all occurrences of the word \"dangerous\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.18, 4.72, "dangerous"]], "response_mode": "offline", "answer": "[{\"word\":\"dangerous\",\"start\":4.18,\"end\":4.72}]"} {"utterance_id": "8555-284449-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0020.flac", "task_type": "lexical_detection", "query": "combined", "prompt": "After the audio ends, extract all occurrences of the word \"combined\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.4, 0.91, "combined"]], "response_mode": "offline", "answer": "[{\"word\":\"combined\",\"start\":0.4,\"end\":0.91}]"} {"utterance_id": "1580-141083-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0007.flac", "task_type": "lexical_detection", "query": "someone", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"someone\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.63, 2.89, "someone"]], "response_mode": "offline", "answer": "[{\"word\":\"someone\",\"start\":2.63,\"end\":2.89}]"} {"utterance_id": "5683-32879-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0002.flac", "task_type": "lexical_detection", "query": "rustling", "prompt": "Once the audio has finished, identify every instance of the word \"rustling\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.61, 3.14, "rustling"]], "response_mode": "offline", "answer": "[{\"word\":\"rustling\",\"start\":2.61,\"end\":3.14}]"} {"utterance_id": "6829-68769-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0001.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.72, 2.4, "mister watson"], [3.48, 4.15, "tom gates"]], "response_mode": "offline", "answer": "[{\"word\":\"mister watson\",\"start\":1.72,\"end\":2.4},{\"word\":\"tom gates\",\"start\":3.48,\"end\":4.15}]"} {"utterance_id": "260-123286-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0027.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "When the utterance is complete, report all instances of an animal name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.78, 2.75, "ichthyosaurus"]], "response_mode": "offline", "answer": "[{\"word\":\"ichthyosaurus\",\"start\":1.78,\"end\":2.75}]"} {"utterance_id": "8455-210777-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0036.flac", "task_type": "lexical_detection", "query": "afford", "prompt": "After listening to the whole recording, output every instance of the word \"afford\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.79, 1.17, "afford"]], "response_mode": "offline", "answer": "[{\"word\":\"afford\",\"start\":0.79,\"end\":1.17}]"} {"utterance_id": "1188-133604-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0005.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Process the full audio offline. Find every instance of a plant name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.23, 2.68, "flower"]], "response_mode": "offline", "answer": "[{\"word\":\"flower\",\"start\":2.23,\"end\":2.68}]"} {"utterance_id": "4992-41806-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0015.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Once the audio has finished, identify every instance of a family relation. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.92, 3.32, "husband"]], "response_mode": "offline", "answer": "[{\"word\":\"husband\",\"start\":2.92,\"end\":3.32}]"} {"utterance_id": "2830-3980-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0032.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "When the utterance is complete, report how a body part accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.81, 2.03, "one"], [3.66, 3.86, "two"], [5.98, 6.27, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.81,\"end\":2.03},{\"word\":\"two\",\"start\":3.66,\"end\":3.86},{\"word\":\"three\",\"start\":5.98,\"end\":6.27}]"} {"utterance_id": "4077-13754-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0003.flac", "task_type": "lexical_detection", "query": "inclined", "prompt": "Process the full audio offline. Find every instance of the word \"inclined\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.65, 2.1, "inclined"]], "response_mode": "offline", "answer": "[{\"word\":\"inclined\",\"start\":1.65,\"end\":2.1}]"} {"utterance_id": "672-122797-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0002.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Once the audio has finished, identify every instance of a family relation. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[7.86, 8.37, "children"]], "response_mode": "offline", "answer": "[{\"word\":\"children\",\"start\":7.86,\"end\":8.37}]"} {"utterance_id": "6829-68771-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0017.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 0.98, "housekeeper"]], "response_mode": "offline", "answer": "[{\"word\":\"housekeeper\",\"start\":0.42,\"end\":0.98}]"} {"utterance_id": "5142-33396-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0014.flac", "task_type": "lexical_detection", "query": "thirty", "prompt": "Once the audio has finished, identify every instance of the word \"thirty\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.18, 0.58, "thirty"]], "response_mode": "offline", "answer": "[{\"word\":\"thirty\",\"start\":0.18,\"end\":0.58}]"} {"utterance_id": "4077-13751-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0009.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, list every instance of an organization name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.38, 2.82, "church"]], "response_mode": "offline", "answer": "[{\"word\":\"church\",\"start\":2.38,\"end\":2.82}]"} {"utterance_id": "260-123288-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0005.flac", "task_type": "lexical_detection", "query": "dense", "prompt": "Process the full audio offline. Find every instance of the word \"dense\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.77, 7.11, "dense"]], "response_mode": "offline", "answer": "[{\"word\":\"dense\",\"start\":6.77,\"end\":7.11}]"} {"utterance_id": "1580-141083-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0015.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[4.04, 4.53, "printer"]], "response_mode": "offline", "answer": "[{\"word\":\"printer\",\"start\":4.04,\"end\":4.53}]"} {"utterance_id": "6930-76324-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0016.flac", "task_type": "lexical_detection", "query": "caught", "prompt": "Process the full audio offline. Find every instance of the word \"caught\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.88, 8.27, "caught"]], "response_mode": "offline", "answer": "[{\"word\":\"caught\",\"start\":7.88,\"end\":8.27}]"} {"utterance_id": "4970-29093-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0010.flac", "task_type": "lexical_detection", "query": "platform", "prompt": "Process the full audio offline. Find every instance of the word \"platform\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.46, 9.12, "platform"]], "response_mode": "offline", "answer": "[{\"word\":\"platform\",\"start\":8.46,\"end\":9.12}]"} {"utterance_id": "1221-135767-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.51, 1.86, "pearl"], [6.79, 7.24, "hester"]], "response_mode": "offline", "answer": "[{\"word\":\"pearl\",\"start\":1.51,\"end\":1.86},{\"word\":\"hester\",\"start\":6.79,\"end\":7.24}]"} {"utterance_id": "7729-102255-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0020.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.96, 7.31, "reeder"]], "response_mode": "offline", "answer": "[{\"word\":\"reeder\",\"start\":6.96,\"end\":7.31}]"} {"utterance_id": "237-134493-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.96, 2.23, "marie emil"]], "response_mode": "offline", "answer": "[{\"word\":\"marie emil\",\"start\":0.96,\"end\":2.23}]"} {"utterance_id": "260-123288-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0009.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "You will hear a complete utterance. When it ends, list every instance of a weather term you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.47, 0.84, "clouds"]], "response_mode": "offline", "answer": "[{\"word\":\"clouds\",\"start\":0.47,\"end\":0.84}]"} {"utterance_id": "672-122797-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0029.flac", "task_type": "lexical_detection", "query": "shine", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"shine\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.87, 1.71, "shine"]], "response_mode": "offline", "answer": "[{\"word\":\"shine\",\"start\":0.87,\"end\":1.71}]"} {"utterance_id": "3729-6852-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0042.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.42, 10.82, "teacher"]], "response_mode": "offline", "answer": "[{\"word\":\"teacher\",\"start\":10.42,\"end\":10.82}]"} {"utterance_id": "6930-76324-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0016.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.62, 7.12, "cynthia's"]], "response_mode": "offline", "answer": "[{\"word\":\"cynthia's\",\"start\":6.62,\"end\":7.12}]"} {"utterance_id": "1089-134686-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0019.flac", "task_type": "lexical_detection", "query": "james", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"james\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.62, 2.08, "james"]], "response_mode": "offline", "answer": "[{\"word\":\"james\",\"start\":1.62,\"end\":2.08}]"} {"utterance_id": "4077-13754-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0000.flac", "task_type": "lexical_detection", "query": "wealth", "prompt": "Once the audio has finished, identify every instance of the word \"wealth\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.98, 4.42, "wealth"]], "response_mode": "offline", "answer": "[{\"word\":\"wealth\",\"start\":3.98,\"end\":4.42}]"} {"utterance_id": "7729-102255-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0006.flac", "task_type": "lexical_detection", "query": "ruffian", "prompt": "When the utterance is complete, report all instances of the word \"ruffian\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.55, 5.01, "ruffian"]], "response_mode": "offline", "answer": "[{\"word\":\"ruffian\",\"start\":4.55,\"end\":5.01}]"} {"utterance_id": "672-122797-0063", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0063.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.03, 2.64, "rats"]], "response_mode": "offline", "answer": "[{\"word\":\"rats\",\"start\":2.03,\"end\":2.64}]"} {"utterance_id": "1580-141084-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0010.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a color name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.85, 1.16, "black"]], "response_mode": "offline", "answer": "[{\"word\":\"black\",\"start\":0.85,\"end\":1.16}]"} {"utterance_id": "61-70968-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0032.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "Once the audio has finished, list each occurrence of a profession as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[1.15, 1.52, "one"], [2.76, 3.19, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.15,\"end\":1.52},{\"word\":\"two\",\"start\":2.76,\"end\":3.19}]"} {"utterance_id": "2830-3980-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0037.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.55, 4.1, "apostle"]], "response_mode": "offline", "answer": "[{\"word\":\"apostle\",\"start\":3.55,\"end\":4.1}]"} {"utterance_id": "4446-2273-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0032.flac", "task_type": "semantic_detection", "query": "color", "prompt": "When the utterance is complete, report all instances of a color name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.25, 6.56, "white"]], "response_mode": "offline", "answer": "[{\"word\":\"white\",\"start\":6.25,\"end\":6.56}]"} {"utterance_id": "4077-13754-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0002.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "When the utterance is complete, report all instances of an organization name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[16.49, 17.43, "secessionists"]], "response_mode": "offline", "answer": "[{\"word\":\"secessionists\",\"start\":16.49,\"end\":17.43}]"} {"utterance_id": "7127-75946-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0021.flac", "task_type": "lexical_detection", "query": "ballet", "prompt": "When the utterance is complete, report all instances of the word \"ballet\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.87, "ballet"]], "response_mode": "offline", "answer": "[{\"word\":\"ballet\",\"start\":2.35,\"end\":2.87}]"} {"utterance_id": "672-122797-0062", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0062.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.95, 2.42, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":1.95,\"end\":2.42}]"} {"utterance_id": "121-127105-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0006.flac", "task_type": "lexical_detection", "query": "others", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"others\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.39, 0.69, "others"]], "response_mode": "offline", "answer": "[{\"word\":\"others\",\"start\":0.39,\"end\":0.69}]"} {"utterance_id": "8224-274381-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0006.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.68, 15.46, "montrose"]], "response_mode": "offline", "answer": "[{\"word\":\"montrose\",\"start\":14.68,\"end\":15.46}]"} {"utterance_id": "1284-1180-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0022.flac", "task_type": "semantic_detection", "query": "country", "prompt": "When the utterance is complete, report all instances of a country name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.32, 2.71, "oz"]], "response_mode": "offline", "answer": "[{\"word\":\"oz\",\"start\":2.32,\"end\":2.71}]"} {"utterance_id": "5142-33396-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0018.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.78, 1.12, "belly"]], "response_mode": "offline", "answer": "[{\"word\":\"belly\",\"start\":0.78,\"end\":1.12}]"} {"utterance_id": "7176-88083-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0013.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a color name with fields word, start, end (seconds). If none, reply [].", "labels": [[8.24, 8.63, "dark"]], "response_mode": "offline", "answer": "[{\"word\":\"dark\",\"start\":8.24,\"end\":8.63}]"} {"utterance_id": "6829-68769-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0031.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen carefully to the entire clip, then report every instance of a money or currency term as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.57, 4.92, "loss"]], "response_mode": "offline", "answer": "[{\"word\":\"loss\",\"start\":4.57,\"end\":4.92}]"} {"utterance_id": "1221-135767-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0016.flac", "task_type": "lexical_detection", "query": "materials", "prompt": "When the utterance is complete, report all instances of the word \"materials\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.49, 4.21, "materials"]], "response_mode": "offline", "answer": "[{\"word\":\"materials\",\"start\":3.49,\"end\":4.21}]"} {"utterance_id": "7176-92135-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0019.flac", "task_type": "lexical_detection", "query": "double", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"double\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.72, 1.01, "double"]], "response_mode": "offline", "answer": "[{\"word\":\"double\",\"start\":0.72,\"end\":1.01}]"} {"utterance_id": "908-157963-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0005.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Once the audio has finished, identify every instance of an animal name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.69, 1.08, "doves"]], "response_mode": "offline", "answer": "[{\"word\":\"doves\",\"start\":0.69,\"end\":1.08}]"} {"utterance_id": "4507-16021-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0003.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Once the audio has finished, identify every instance of a family relation. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.06, 1.59, "son"], [3.22, 3.79, "daughter"]], "response_mode": "offline", "answer": "[{\"word\":\"son\",\"start\":1.06,\"end\":1.59},{\"word\":\"daughter\",\"start\":3.22,\"end\":3.79}]"} {"utterance_id": "4077-13751-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0014.flac", "task_type": "lexical_detection", "query": "afterward", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"afterward\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.04, 7.5, "afterward"]], "response_mode": "offline", "answer": "[{\"word\":\"afterward\",\"start\":7.04,\"end\":7.5}]"} {"utterance_id": "237-134493-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0013.flac", "task_type": "lexical_detection", "query": "purpose", "prompt": "When the utterance is complete, report all instances of the word \"purpose\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.25, 2.75, "purpose"]], "response_mode": "offline", "answer": "[{\"word\":\"purpose\",\"start\":2.25,\"end\":2.75}]"} {"utterance_id": "4077-13754-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0012.flac", "task_type": "lexical_counting", "query": "dollars", "prompt": "After listening to the whole recording, output timed count words for the word \"dollars\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[9.22, 9.77, "one"], [21.13, 21.75, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":9.22,\"end\":9.77},{\"word\":\"two\",\"start\":21.13,\"end\":21.75}]"} {"utterance_id": "5142-33396-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0030.flac", "task_type": "lexical_detection", "query": "thralls", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"thralls\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 0.83, "thralls"]], "response_mode": "offline", "answer": "[{\"word\":\"thralls\",\"start\":0.45,\"end\":0.83}]"} {"utterance_id": "672-122797-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0011.flac", "task_type": "lexical_detection", "query": "happens", "prompt": "Process the full audio offline. Find every instance of the word \"happens\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.34, 1.83, "happens"]], "response_mode": "offline", "answer": "[{\"word\":\"happens\",\"start\":1.34,\"end\":1.83}]"} {"utterance_id": "4446-2271-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0006.flac", "task_type": "lexical_detection", "query": "wanting", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"wanting\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.53, 0.76, "wanting"]], "response_mode": "offline", "answer": "[{\"word\":\"wanting\",\"start\":0.53,\"end\":0.76}]"} {"utterance_id": "5142-36377-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0015.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.26, 0.54, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":0.26,\"end\":0.54}]"} {"utterance_id": "3575-170457-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0012.flac", "task_type": "lexical_detection", "query": "need", "prompt": "After listening to the whole recording, output every instance of the word \"need\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.99, 2.21, "need"]], "response_mode": "offline", "answer": "[{\"word\":\"need\",\"start\":1.99,\"end\":2.21}]"} {"utterance_id": "1188-133604-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0008.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Once the audio has finished, identify every instance of a color name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.88, 7.34, "red"], [9.41, 9.92, "green"]], "response_mode": "offline", "answer": "[{\"word\":\"red\",\"start\":6.88,\"end\":7.34},{\"word\":\"green\",\"start\":9.41,\"end\":9.92}]"} {"utterance_id": "1221-135766-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0000.flac", "task_type": "lexical_detection", "query": "quivering", "prompt": "When the utterance is complete, report all instances of the word \"quivering\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.66, 9.14, "quivering"]], "response_mode": "offline", "answer": "[{\"word\":\"quivering\",\"start\":8.66,\"end\":9.14}]"} {"utterance_id": "1284-1181-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0014.flac", "task_type": "lexical_counting", "query": "small", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"small\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[1.14, 1.54, "one"], [6.88, 7.21, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.14,\"end\":1.54},{\"word\":\"two\",\"start\":6.88,\"end\":7.21}]"} {"utterance_id": "1580-141084-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0044.flac", "task_type": "lexical_detection", "query": "gloves", "prompt": "Once the audio has finished, identify every instance of the word \"gloves\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.41, 1.19, "gloves"]], "response_mode": "offline", "answer": "[{\"word\":\"gloves\",\"start\":0.41,\"end\":1.19}]"} {"utterance_id": "1221-135766-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0004.flac", "task_type": "lexical_detection", "query": "inner", "prompt": "When the utterance is complete, report all instances of the word \"inner\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.48, 6.76, "inner"]], "response_mode": "offline", "answer": "[{\"word\":\"inner\",\"start\":6.48,\"end\":6.76}]"} {"utterance_id": "6829-68771-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0016.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.43, 3.93, "aunt jane"], [5.39, 5.66, "beth"]], "response_mode": "offline", "answer": "[{\"word\":\"aunt jane\",\"start\":3.43,\"end\":3.93},{\"word\":\"beth\",\"start\":5.39,\"end\":5.66}]"} {"utterance_id": "5683-32879-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0012.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the full audio. After it finishes, find every instance of a family relation. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.92, 2.44, "cousin"]], "response_mode": "offline", "answer": "[{\"word\":\"cousin\",\"start\":1.92,\"end\":2.44}]"} {"utterance_id": "8230-279154-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0040.flac", "task_type": "lexical_counting", "query": "recognition", "prompt": "When the utterance is complete, report how the word \"recognition\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.55, 5.3, "one"], [10.67, 11.4, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.55,\"end\":5.3},{\"word\":\"two\",\"start\":10.67,\"end\":11.4}]"} {"utterance_id": "2961-960-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0013.flac", "task_type": "lexical_detection", "query": "probable", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"probable\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.82, 1.41, "probable"]], "response_mode": "offline", "answer": "[{\"word\":\"probable\",\"start\":0.82,\"end\":1.41}]"} {"utterance_id": "1284-1180-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0000.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.37, 2.58, "knee"], [5.02, 5.46, "waist"]], "response_mode": "offline", "answer": "[{\"word\":\"knee\",\"start\":2.37,\"end\":2.58},{\"word\":\"waist\",\"start\":5.02,\"end\":5.46}]"} {"utterance_id": "7729-102255-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0006.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an ordinal number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.48, 3.87, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":3.48,\"end\":3.87}]"} {"utterance_id": "5105-28233-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28233/5105-28233-0005.flac", "task_type": "lexical_counting", "query": "sometimes", "prompt": "After listening to the whole recording, output timed count words for the word \"sometimes\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[0.42, 0.94, "one"], [3.82, 4.36, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.42,\"end\":0.94},{\"word\":\"two\",\"start\":3.82,\"end\":4.36}]"} {"utterance_id": "2961-960-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0018.flac", "task_type": "lexical_detection", "query": "dramatic", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"dramatic\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.44, 6.07, "dramatic"]], "response_mode": "offline", "answer": "[{\"word\":\"dramatic\",\"start\":5.44,\"end\":6.07}]"} {"utterance_id": "2094-142345-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0032.flac", "task_type": "lexical_detection", "query": "sort", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"sort\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.23, 2.47, "sort"]], "response_mode": "offline", "answer": "[{\"word\":\"sort\",\"start\":2.23,\"end\":2.47}]"} {"utterance_id": "5639-40744-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0006.flac", "task_type": "lexical_detection", "query": "parents", "prompt": "Process the full audio offline. Find every instance of the word \"parents\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.06, 5.61, "parents"]], "response_mode": "offline", "answer": "[{\"word\":\"parents\",\"start\":5.06,\"end\":5.61}]"} {"utterance_id": "7729-102255-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0027.flac", "task_type": "lexical_detection", "query": "footnote", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"footnote\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.56, 1.22, "footnote"]], "response_mode": "offline", "answer": "[{\"word\":\"footnote\",\"start\":0.56,\"end\":1.22}]"} {"utterance_id": "1320-122617-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0033.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.98, 9.49, "warriors"]], "response_mode": "offline", "answer": "[{\"word\":\"warriors\",\"start\":8.98,\"end\":9.49}]"} {"utterance_id": "5639-40744-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0021.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.81, 8.09, "eye"], [10.07, 10.57, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"eye\",\"start\":7.81,\"end\":8.09},{\"word\":\"face\",\"start\":10.07,\"end\":10.57}]"} {"utterance_id": "260-123286-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0011.flac", "task_type": "semantic_counting", "query": "weather", "prompt": "After the audio ends, tally every instance of a weather term. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[1.23, 1.56, "one"], [2.92, 3.25, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.23,\"end\":1.56},{\"word\":\"two\",\"start\":2.92,\"end\":3.25}]"} {"utterance_id": "5105-28240-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0001.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[3.71, 4.01, "six"]], "response_mode": "offline", "answer": "[{\"word\":\"six\",\"start\":3.71,\"end\":4.01}]"} {"utterance_id": "2830-3980-0076", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0076.flac", "task_type": "lexical_detection", "query": "despair", "prompt": "Once the audio has finished, identify every instance of the word \"despair\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.81, 4.4, "despair"]], "response_mode": "offline", "answer": "[{\"word\":\"despair\",\"start\":3.81,\"end\":4.4}]"} {"utterance_id": "4970-29093-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0006.flac", "task_type": "lexical_counting", "query": "settle", "prompt": "Once the audio has finished, list each occurrence of the word \"settle\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[16.06, 16.5, "one"], [18.65, 19.25, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":16.06,\"end\":16.5},{\"word\":\"two\",\"start\":18.65,\"end\":19.25}]"} {"utterance_id": "2830-3980-0075", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0075.flac", "task_type": "lexical_detection", "query": "significance", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"significance\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.98, 1.61, "significance"]], "response_mode": "offline", "answer": "[{\"word\":\"significance\",\"start\":0.98,\"end\":1.61}]"} {"utterance_id": "61-70968-0056", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0056.flac", "task_type": "semantic_detection", "query": "food", "prompt": "When the utterance is complete, report all instances of a food name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 0.75, "wine"]], "response_mode": "offline", "answer": "[{\"word\":\"wine\",\"start\":0.52,\"end\":0.75}]"} {"utterance_id": "2300-131720-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0019.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.09, 2.48, "edison"]], "response_mode": "offline", "answer": "[{\"word\":\"edison\",\"start\":2.09,\"end\":2.48}]"} {"utterance_id": "7176-92135-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0039.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.87, 2.32, "butler"]], "response_mode": "offline", "answer": "[{\"word\":\"butler\",\"start\":1.87,\"end\":2.32}]"} {"utterance_id": "5142-36377-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0024.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.14, 0.52, "naomi"]], "response_mode": "offline", "answer": "[{\"word\":\"naomi\",\"start\":0.14,\"end\":0.52}]"} {"utterance_id": "2830-3980-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0034.flac", "task_type": "lexical_detection", "query": "cannot", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"cannot\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.34, 1.69, "cannot"]], "response_mode": "offline", "answer": "[{\"word\":\"cannot\",\"start\":1.34,\"end\":1.69}]"} {"utterance_id": "4446-2275-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0014.flac", "task_type": "lexical_detection", "query": "miserable", "prompt": "When the utterance is complete, report all instances of the word \"miserable\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.43, 2.03, "miserable"]], "response_mode": "offline", "answer": "[{\"word\":\"miserable\",\"start\":1.43,\"end\":2.03}]"} {"utterance_id": "3575-170457-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0036.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.73, 1.05, "eyes"], [19.66, 20.22, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":0.73,\"end\":1.05},{\"word\":\"heart\",\"start\":19.66,\"end\":20.22}]"} {"utterance_id": "260-123288-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0011.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "You will hear a complete utterance. When it ends, list every instance of a weather term you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.36, 6.91, "tempest"]], "response_mode": "offline", "answer": "[{\"word\":\"tempest\",\"start\":6.36,\"end\":6.91}]"} {"utterance_id": "3570-5694-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0001.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen carefully to the entire clip, then report every instance of a money or currency term as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.71, 3.17, "wealth"]], "response_mode": "offline", "answer": "[{\"word\":\"wealth\",\"start\":2.71,\"end\":3.17}]"} {"utterance_id": "3575-170457-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0030.flac", "task_type": "lexical_detection", "query": "contained", "prompt": "When the utterance is complete, report all instances of the word \"contained\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.95, 4.44, "contained"]], "response_mode": "offline", "answer": "[{\"word\":\"contained\",\"start\":3.95,\"end\":4.44}]"} {"utterance_id": "260-123288-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0001.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Once the audio has finished, identify every instance of a weather term. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.71, 1.16, "weather"]], "response_mode": "offline", "answer": "[{\"word\":\"weather\",\"start\":0.71,\"end\":1.16}]"} {"utterance_id": "3570-5694-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0002.flac", "task_type": "lexical_detection", "query": "directed", "prompt": "When the utterance is complete, report all instances of the word \"directed\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.54, 8.04, "directed"]], "response_mode": "offline", "answer": "[{\"word\":\"directed\",\"start\":7.54,\"end\":8.04}]"} {"utterance_id": "2094-142345-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0052.flac", "task_type": "lexical_detection", "query": "horses", "prompt": "After listening to the whole recording, output every instance of the word \"horses\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.53, 6.26, "horses"]], "response_mode": "offline", "answer": "[{\"word\":\"horses\",\"start\":5.53,\"end\":6.26}]"} {"utterance_id": "5142-33396-0055", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0055.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the full audio. After it finishes, find every instance of a family relation. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.64, 2.06, "father's"]], "response_mode": "offline", "answer": "[{\"word\":\"father's\",\"start\":1.64,\"end\":2.06}]"} {"utterance_id": "8230-279154-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0037.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "Once the audio has finished, list each occurrence of an animal name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[11.53, 12.09, "one"], [18.81, 19.28, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":11.53,\"end\":12.09},{\"word\":\"two\",\"start\":18.81,\"end\":19.28}]"} {"utterance_id": "8463-287645-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0014.flac", "task_type": "lexical_detection", "query": "didn't", "prompt": "Process the full audio offline. Find every instance of the word \"didn't\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.48, 1.74, "didn't"]], "response_mode": "offline", "answer": "[{\"word\":\"didn't\",\"start\":1.48,\"end\":1.74}]"} {"utterance_id": "1188-133604-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0026.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After the audio ends, extract all occurrences of a plant name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.31, 6.74, "trees"], [8.9, 9.43, "trees"]], "response_mode": "offline", "answer": "[{\"word\":\"trees\",\"start\":6.31,\"end\":6.74},{\"word\":\"trees\",\"start\":8.9,\"end\":9.43}]"} {"utterance_id": "1995-1837-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0018.flac", "task_type": "lexical_detection", "query": "creek", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"creek\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.36, 6.79, "creek"]], "response_mode": "offline", "answer": "[{\"word\":\"creek\",\"start\":6.36,\"end\":6.79}]"} {"utterance_id": "2300-131720-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0026.flac", "task_type": "lexical_counting", "query": "lamp", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"lamp\" with word/start/end. If none, reply [].", "labels": [[0.72, 1.16, "one"], [15.53, 15.92, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.72,\"end\":1.16},{\"word\":\"two\",\"start\":15.53,\"end\":15.92}]"} {"utterance_id": "2300-131720-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0023.flac", "task_type": "lexical_detection", "query": "fourth", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"fourth\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.11, 9.57, "fourth"]], "response_mode": "offline", "answer": "[{\"word\":\"fourth\",\"start\":9.11,\"end\":9.57}]"} {"utterance_id": "7127-75947-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0034.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.12, 1.43, "king"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":1.12,\"end\":1.43}]"} {"utterance_id": "4446-2271-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0017.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Process the full audio offline. For every occurrence of a number, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.35, 0.6, "one"], [7.92, 8.2, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.35,\"end\":0.6},{\"word\":\"two\",\"start\":7.92,\"end\":8.2}]"} {"utterance_id": "8463-294825-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0014.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.1, 1.51, "six"]], "response_mode": "offline", "answer": "[{\"word\":\"six\",\"start\":1.1,\"end\":1.51}]"} {"utterance_id": "3729-6852-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0017.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "When the utterance is complete, report how a plant name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.12, 7.73, "one"], [16.79, 17.47, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.12,\"end\":7.73},{\"word\":\"two\",\"start\":16.79,\"end\":17.47}]"} {"utterance_id": "1580-141083-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0034.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 0.84, "holmes"]], "response_mode": "offline", "answer": "[{\"word\":\"holmes\",\"start\":0.51,\"end\":0.84}]"} {"utterance_id": "7021-79730-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.96, 5.53, "mary"]], "response_mode": "offline", "answer": "[{\"word\":\"mary\",\"start\":4.96,\"end\":5.53}]"} {"utterance_id": "2094-142345-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0029.flac", "task_type": "lexical_detection", "query": "anybody", "prompt": "After the audio ends, extract all occurrences of the word \"anybody\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.9, 1.43, "anybody"]], "response_mode": "offline", "answer": "[{\"word\":\"anybody\",\"start\":0.9,\"end\":1.43}]"} {"utterance_id": "3575-170457-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0052.flac", "task_type": "lexical_detection", "query": "weight", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"weight\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.2, 1.52, "weight"]], "response_mode": "offline", "answer": "[{\"word\":\"weight\",\"start\":1.2,\"end\":1.52}]"} {"utterance_id": "237-126133-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0014.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.94, 4.4, "arms"], [6.09, 6.58, "side"]], "response_mode": "offline", "answer": "[{\"word\":\"arms\",\"start\":3.94,\"end\":4.4},{\"word\":\"side\",\"start\":6.09,\"end\":6.58}]"} {"utterance_id": "8224-274381-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0007.flac", "task_type": "lexical_detection", "query": "forces", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"forces\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.7, 8.29, "forces"]], "response_mode": "offline", "answer": "[{\"word\":\"forces\",\"start\":7.7,\"end\":8.29}]"} {"utterance_id": "5142-36586-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36586/5142-36586-0004.flac", "task_type": "lexical_detection", "query": "effects", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"effects\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.95, "effects"]], "response_mode": "offline", "answer": "[{\"word\":\"effects\",\"start\":0.53,\"end\":0.95}]"} {"utterance_id": "8555-284447-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0008.flac", "task_type": "lexical_detection", "query": "glittered", "prompt": "Process the full audio offline. Find every instance of the word \"glittered\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.22, 2.58, "glittered"]], "response_mode": "offline", "answer": "[{\"word\":\"glittered\",\"start\":2.22,\"end\":2.58}]"} {"utterance_id": "2094-142345-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0008.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "After the audio ends, extract all occurrences of a weather term. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.87, 5.31, "rain"]], "response_mode": "offline", "answer": "[{\"word\":\"rain\",\"start\":4.87,\"end\":5.31}]"} {"utterance_id": "7729-102255-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0016.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.33, 0.82, "governor"]], "response_mode": "offline", "answer": "[{\"word\":\"governor\",\"start\":0.33,\"end\":0.82}]"} {"utterance_id": "3575-170457-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0005.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.92, 2.27, "daughter"]], "response_mode": "offline", "answer": "[{\"word\":\"daughter\",\"start\":1.92,\"end\":2.27}]"} {"utterance_id": "4077-13751-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0003.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.09, 14.69, "officers"]], "response_mode": "offline", "answer": "[{\"word\":\"officers\",\"start\":14.09,\"end\":14.69}]"} {"utterance_id": "8463-294828-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0023.flac", "task_type": "lexical_detection", "query": "friend", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"friend\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.03, 1.39, "friend"]], "response_mode": "offline", "answer": "[{\"word\":\"friend\",\"start\":1.03,\"end\":1.39}]"} {"utterance_id": "4970-29095-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.87, 2.33, "abner"]], "response_mode": "offline", "answer": "[{\"word\":\"abner\",\"start\":1.87,\"end\":2.33}]"} {"utterance_id": "7021-85628-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0007.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.45, 6.91, "soldiers"], [11.9, 12.62, "courtiers"]], "response_mode": "offline", "answer": "[{\"word\":\"soldiers\",\"start\":6.45,\"end\":6.91},{\"word\":\"courtiers\",\"start\":11.9,\"end\":12.62}]"} {"utterance_id": "6930-75918-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0006.flac", "task_type": "lexical_detection", "query": "fixed", "prompt": "When the utterance is complete, report all instances of the word \"fixed\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.53, 4.89, "fixed"]], "response_mode": "offline", "answer": "[{\"word\":\"fixed\",\"start\":4.53,\"end\":4.89}]"} {"utterance_id": "8555-292519-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0005.flac", "task_type": "lexical_detection", "query": "gleaming", "prompt": "After listening to the whole recording, output every instance of the word \"gleaming\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.35, 4.88, "gleaming"]], "response_mode": "offline", "answer": "[{\"word\":\"gleaming\",\"start\":4.35,\"end\":4.88}]"} {"utterance_id": "7021-79759-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79759/7021-79759-0002.flac", "task_type": "lexical_detection", "query": "chiefly", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"chiefly\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.59, 1.02, "chiefly"]], "response_mode": "offline", "answer": "[{\"word\":\"chiefly\",\"start\":0.59,\"end\":1.02}]"} {"utterance_id": "5142-33396-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0044.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.23, 4.78, "farmer"]], "response_mode": "offline", "answer": "[{\"word\":\"farmer\",\"start\":4.23,\"end\":4.78}]"} {"utterance_id": "6829-68771-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0014.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.5, 2.08, "attendants"]], "response_mode": "offline", "answer": "[{\"word\":\"attendants\",\"start\":1.5,\"end\":2.08}]"} {"utterance_id": "5142-33396-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0052.flac", "task_type": "lexical_detection", "query": "ring", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"ring\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.57, 0.86, "ring"]], "response_mode": "offline", "answer": "[{\"word\":\"ring\",\"start\":0.57,\"end\":0.86}]"} {"utterance_id": "7021-85628-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0024.flac", "task_type": "lexical_detection", "query": "red", "prompt": "After listening to the whole recording, output every instance of the word \"red\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.46, 1.9, "red"]], "response_mode": "offline", "answer": "[{\"word\":\"red\",\"start\":1.46,\"end\":1.9}]"} {"utterance_id": "4992-41797-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0021.flac", "task_type": "semantic_detection", "query": "city", "prompt": "After listening to the whole recording, output every instance of a city name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.16, 8.01, "charlestown"]], "response_mode": "offline", "answer": "[{\"word\":\"charlestown\",\"start\":7.16,\"end\":8.01}]"} {"utterance_id": "1221-135766-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0006.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.31, 3.79, "child's"]], "response_mode": "offline", "answer": "[{\"word\":\"child's\",\"start\":3.31,\"end\":3.79}]"} {"utterance_id": "7021-79740-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.76, 1.42, "delia"]], "response_mode": "offline", "answer": "[{\"word\":\"delia\",\"start\":0.76,\"end\":1.42}]"} {"utterance_id": "6829-68769-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0028.flac", "task_type": "lexical_detection", "query": "supposed", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"supposed\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.63, 1.04, "supposed"]], "response_mode": "offline", "answer": "[{\"word\":\"supposed\",\"start\":0.63,\"end\":1.04}]"} {"utterance_id": "1188-133604-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0008.flac", "task_type": "semantic_counting", "query": "color", "prompt": "Once the audio has finished, list each occurrence of a color name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[6.88, 7.34, "one"], [9.41, 9.92, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.88,\"end\":7.34},{\"word\":\"two\",\"start\":9.41,\"end\":9.92}]"} {"utterance_id": "3570-5694-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0021.flac", "task_type": "lexical_detection", "query": "employments", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"employments\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.21, 5.78, "employments"]], "response_mode": "offline", "answer": "[{\"word\":\"employments\",\"start\":5.21,\"end\":5.78}]"} {"utterance_id": "237-134500-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0002.flac", "task_type": "semantic_counting", "query": "weather", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a weather term with word/start/end. If none, reply [].", "labels": [[0.76, 1.05, "one"], [3.26, 3.71, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.76,\"end\":1.05},{\"word\":\"two\",\"start\":3.26,\"end\":3.71}]"} {"utterance_id": "4446-2275-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0017.flac", "task_type": "lexical_detection", "query": "unbearable", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"unbearable\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.54, 2.23, "unbearable"]], "response_mode": "offline", "answer": "[{\"word\":\"unbearable\",\"start\":1.54,\"end\":2.23}]"} {"utterance_id": "8555-292519-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0000.flac", "task_type": "lexical_detection", "query": "break", "prompt": "When the utterance is complete, report all instances of the word \"break\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.59, 10.94, "break"]], "response_mode": "offline", "answer": "[{\"word\":\"break\",\"start\":10.59,\"end\":10.94}]"} {"utterance_id": "1580-141083-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0017.flac", "task_type": "lexical_detection", "query": "seems", "prompt": "After listening to the whole recording, output every instance of the word \"seems\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.89, 1.22, "seems"]], "response_mode": "offline", "answer": "[{\"word\":\"seems\",\"start\":0.89,\"end\":1.22}]"} {"utterance_id": "7021-85628-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0015.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Process the full audio offline. Find every instance of a color name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.19, 7.64, "gold"]], "response_mode": "offline", "answer": "[{\"word\":\"gold\",\"start\":7.19,\"end\":7.64}]"} {"utterance_id": "7127-75947-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0007.flac", "task_type": "lexical_detection", "query": "dance", "prompt": "Once the audio has finished, identify every instance of the word \"dance\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.55, 5.09, "dance"]], "response_mode": "offline", "answer": "[{\"word\":\"dance\",\"start\":4.55,\"end\":5.09}]"} {"utterance_id": "908-31957-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0002.flac", "task_type": "lexical_detection", "query": "placed", "prompt": "After the audio ends, extract all occurrences of the word \"placed\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.79, 3.35, "placed"]], "response_mode": "offline", "answer": "[{\"word\":\"placed\",\"start\":2.79,\"end\":3.35}]"} {"utterance_id": "6829-68771-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0012.flac", "task_type": "lexical_counting", "query": "within", "prompt": "After the audio ends, tally every instance of the word \"within\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[1.89, 2.37, "one"], [7.68, 8.02, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.89,\"end\":2.37},{\"word\":\"two\",\"start\":7.68,\"end\":8.02}]"} {"utterance_id": "121-123852-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123852/121-123852-0004.flac", "task_type": "lexical_detection", "query": "doth", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"doth\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.02, 1.41, "doth"], [10.04, 10.45, "doth"]], "response_mode": "offline", "answer": "[{\"word\":\"doth\",\"start\":1.02,\"end\":1.41},{\"word\":\"doth\",\"start\":10.04,\"end\":10.45}]"} {"utterance_id": "7176-88083-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0009.flac", "task_type": "lexical_detection", "query": "prey", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"prey\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.66, 3.01, "prey"]], "response_mode": "offline", "answer": "[{\"word\":\"prey\",\"start\":2.66,\"end\":3.01}]"} {"utterance_id": "6829-68769-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0024.flac", "task_type": "lexical_detection", "query": "helpless", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"helpless\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.31, 2.77, "helpless"]], "response_mode": "offline", "answer": "[{\"word\":\"helpless\",\"start\":2.31,\"end\":2.77}]"} {"utterance_id": "6930-75918-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0015.flac", "task_type": "lexical_detection", "query": "country's", "prompt": "Once the audio has finished, identify every instance of the word \"country's\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.91, 4.47, "country's"]], "response_mode": "offline", "answer": "[{\"word\":\"country's\",\"start\":3.91,\"end\":4.47}]"} {"utterance_id": "7729-102255-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0031.flac", "task_type": "lexical_detection", "query": "search", "prompt": "After the audio ends, extract all occurrences of the word \"search\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.8, 2.36, "search"]], "response_mode": "offline", "answer": "[{\"word\":\"search\",\"start\":1.8,\"end\":2.36}]"} {"utterance_id": "260-123288-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0020.flac", "task_type": "lexical_detection", "query": "raft", "prompt": "When the utterance is complete, report all instances of the word \"raft\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.22, 2.72, "raft"]], "response_mode": "offline", "answer": "[{\"word\":\"raft\",\"start\":2.22,\"end\":2.72}]"} {"utterance_id": "4970-29095-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0005.flac", "task_type": "lexical_detection", "query": "greatly", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"greatly\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.24, 1.72, "greatly"]], "response_mode": "offline", "answer": "[{\"word\":\"greatly\",\"start\":1.24,\"end\":1.72}]"} {"utterance_id": "1284-1181-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0013.flac", "task_type": "lexical_detection", "query": "dared", "prompt": "Process the full audio offline. Find every instance of the word \"dared\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.63, 6.9, "dared"]], "response_mode": "offline", "answer": "[{\"word\":\"dared\",\"start\":6.63,\"end\":6.9}]"} {"utterance_id": "3570-5695-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0011.flac", "task_type": "lexical_detection", "query": "compared", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"compared\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.91, 7.39, "compared"]], "response_mode": "offline", "answer": "[{\"word\":\"compared\",\"start\":6.91,\"end\":7.39}]"} {"utterance_id": "1580-141084-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0032.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Once the audio has finished, list each occurrence of a body part as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[2.21, 2.55, "one"], [4.04, 4.67, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.21,\"end\":2.55},{\"word\":\"two\",\"start\":4.04,\"end\":4.67}]"} {"utterance_id": "4446-2275-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0016.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the full audio. After it finishes, find every instance of a color name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.0, 4.27, "dark"]], "response_mode": "offline", "answer": "[{\"word\":\"dark\",\"start\":4.0,\"end\":4.27}]"} {"utterance_id": "7021-85628-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0011.flac", "task_type": "semantic_counting", "query": "color", "prompt": "After listening to the whole recording, output timed count words for a color name (one, two, three, ...) as JSON. If none, reply [].", "labels": [[2.03, 2.28, "one"], [6.88, 7.4, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.03,\"end\":2.28},{\"word\":\"two\",\"start\":6.88,\"end\":7.4}]"} {"utterance_id": "4970-29093-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0004.flac", "task_type": "lexical_detection", "query": "decide", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"decide\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.4, 1.88, "decide"]], "response_mode": "offline", "answer": "[{\"word\":\"decide\",\"start\":1.4,\"end\":1.88}]"} {"utterance_id": "7176-92135-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0011.flac", "task_type": "lexical_detection", "query": "object", "prompt": "Process the full audio offline. Find every instance of the word \"object\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.71, 1.12, "object"]], "response_mode": "offline", "answer": "[{\"word\":\"object\",\"start\":0.71,\"end\":1.12}]"} {"utterance_id": "1320-122617-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0036.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.89, 8.46, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":7.89,\"end\":8.46}]"} {"utterance_id": "2300-131720-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0001.flac", "task_type": "lexical_detection", "query": "temporary", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"temporary\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.97, 2.69, "temporary"]], "response_mode": "offline", "answer": "[{\"word\":\"temporary\",\"start\":1.97,\"end\":2.69}]"} {"utterance_id": "61-70968-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0016.flac", "task_type": "lexical_detection", "query": "purse", "prompt": "When the utterance is complete, report all instances of the word \"purse\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.8, 3.08, "purse"]], "response_mode": "offline", "answer": "[{\"word\":\"purse\",\"start\":2.8,\"end\":3.08}]"} {"utterance_id": "3729-6852-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0012.flac", "task_type": "lexical_detection", "query": "call", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"call\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.0, 1.3, "call"]], "response_mode": "offline", "answer": "[{\"word\":\"call\",\"start\":1.0,\"end\":1.3}]"} {"utterance_id": "260-123440-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0004.flac", "task_type": "lexical_detection", "query": "fan", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"fan\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.23, 1.72, "fan"]], "response_mode": "offline", "answer": "[{\"word\":\"fan\",\"start\":1.23,\"end\":1.72}]"} {"utterance_id": "4970-29095-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0019.flac", "task_type": "lexical_detection", "query": "broad", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"broad\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.65, 2.05, "broad"]], "response_mode": "offline", "answer": "[{\"word\":\"broad\",\"start\":1.65,\"end\":2.05}]"} {"utterance_id": "672-122797-0074", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0074.flac", "task_type": "lexical_detection", "query": "gone", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"gone\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.86, 5.35, "gone"]], "response_mode": "offline", "answer": "[{\"word\":\"gone\",\"start\":4.86,\"end\":5.35}]"} {"utterance_id": "672-122797-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0051.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.6, 4.35, "fir tree"]], "response_mode": "offline", "answer": "[{\"word\":\"fir tree\",\"start\":3.6,\"end\":4.35}]"} {"utterance_id": "7729-102255-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0009.flac", "task_type": "lexical_detection", "query": "dissent", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"dissent\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 0.98, "dissent"]], "response_mode": "offline", "answer": "[{\"word\":\"dissent\",\"start\":0.45,\"end\":0.98}]"} {"utterance_id": "672-122797-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0008.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "When the utterance is complete, report how a plant name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.63, 4.35, "one"], [11.02, 11.66, "two"], [17.78, 18.37, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.63,\"end\":4.35},{\"word\":\"two\",\"start\":11.02,\"end\":11.66},{\"word\":\"three\",\"start\":17.78,\"end\":18.37}]"} {"utterance_id": "672-122797-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0059.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After the audio ends, extract all occurrences of a plant name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.82, 3.24, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":2.82,\"end\":3.24}]"} {"utterance_id": "1320-122612-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0014.flac", "task_type": "lexical_detection", "query": "discovery", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"discovery\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.52, 3.24, "discovery"]], "response_mode": "offline", "answer": "[{\"word\":\"discovery\",\"start\":2.52,\"end\":3.24}]"} {"utterance_id": "7127-75947-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0018.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.04, 3.65, "la valliere"]], "response_mode": "offline", "answer": "[{\"word\":\"la valliere\",\"start\":3.04,\"end\":3.65}]"} {"utterance_id": "8555-292519-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0009.flac", "task_type": "lexical_detection", "query": "hit", "prompt": "After the audio ends, extract all occurrences of the word \"hit\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.05, 11.32, "hit"]], "response_mode": "offline", "answer": "[{\"word\":\"hit\",\"start\":11.05,\"end\":11.32}]"} {"utterance_id": "1580-141083-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0026.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.33, 2.77, "soames"]], "response_mode": "offline", "answer": "[{\"word\":\"soames\",\"start\":2.33,\"end\":2.77}]"} {"utterance_id": "8463-294828-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0001.flac", "task_type": "lexical_detection", "query": "seconds", "prompt": "When the utterance is complete, report all instances of the word \"seconds\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.84, 1.32, "seconds"]], "response_mode": "offline", "answer": "[{\"word\":\"seconds\",\"start\":0.84,\"end\":1.32}]"} {"utterance_id": "7176-88083-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0007.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Listen to the full audio. Count every occurrence of a body part. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.4, 3.86, "one"], [7.56, 7.97, "two"], [9.4, 9.76, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.4,\"end\":3.86},{\"word\":\"two\",\"start\":7.56,\"end\":7.97},{\"word\":\"three\",\"start\":9.4,\"end\":9.76}]"} {"utterance_id": "8463-294828-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0001.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "After listening to the whole recording, output every instance of an animal name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.94, 6.69, "unicorn"]], "response_mode": "offline", "answer": "[{\"word\":\"unicorn\",\"start\":5.94,\"end\":6.69}]"} {"utterance_id": "61-70968-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0027.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.44, 5.63, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":5.44,\"end\":5.63}]"} {"utterance_id": "5639-40744-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0030.flac", "task_type": "lexical_detection", "query": "blessed", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"blessed\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[16.64, 17.05, "blessed"]], "response_mode": "offline", "answer": "[{\"word\":\"blessed\",\"start\":16.64,\"end\":17.05}]"} {"utterance_id": "1995-1837-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0000.flac", "task_type": "lexical_detection", "query": "silver", "prompt": "When the utterance is complete, report all instances of the word \"silver\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.64, 1.04, "silver"]], "response_mode": "offline", "answer": "[{\"word\":\"silver\",\"start\":0.64,\"end\":1.04}]"} {"utterance_id": "2830-3980-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0054.flac", "task_type": "lexical_detection", "query": "apparent", "prompt": "Process the full audio offline. Find every instance of the word \"apparent\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.43, 2.91, "apparent"]], "response_mode": "offline", "answer": "[{\"word\":\"apparent\",\"start\":2.43,\"end\":2.91}]"} {"utterance_id": "1188-133604-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0026.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "After the audio ends, extract all occurrences of an animal name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[18.9, 19.59, "python"]], "response_mode": "offline", "answer": "[{\"word\":\"python\",\"start\":18.9,\"end\":19.59}]"} {"utterance_id": "7127-75946-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0022.flac", "task_type": "lexical_detection", "query": "devoted", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"devoted\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.18, 3.69, "devoted"]], "response_mode": "offline", "answer": "[{\"word\":\"devoted\",\"start\":3.18,\"end\":3.69}]"} {"utterance_id": "4992-23283-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0018.flac", "task_type": "lexical_detection", "query": "relieve", "prompt": "Process the full audio offline. Find every instance of the word \"relieve\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.43, 0.77, "relieve"]], "response_mode": "offline", "answer": "[{\"word\":\"relieve\",\"start\":0.43,\"end\":0.77}]"} {"utterance_id": "1221-135767-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0005.flac", "task_type": "semantic_counting", "query": "color", "prompt": "When the utterance is complete, report how a color name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.81, 1.27, "one"], [3.28, 3.74, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.81,\"end\":1.27},{\"word\":\"two\",\"start\":3.28,\"end\":3.74}]"} {"utterance_id": "7176-88083-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0011.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "After listening to the whole recording, output timed count words for an animal name (one, two, three, ...) as JSON. If none, reply [].", "labels": [[3.23, 3.58, "one"], [7.87, 8.29, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.23,\"end\":3.58},{\"word\":\"two\",\"start\":7.87,\"end\":8.29}]"} {"utterance_id": "1995-1826-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0011.flac", "task_type": "lexical_detection", "query": "teaching", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"teaching\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.01, 1.42, "teaching"]], "response_mode": "offline", "answer": "[{\"word\":\"teaching\",\"start\":1.01,\"end\":1.42}]"} {"utterance_id": "8555-284447-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0005.flac", "task_type": "lexical_detection", "query": "benches", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"benches\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.43, 4.92, "benches"]], "response_mode": "offline", "answer": "[{\"word\":\"benches\",\"start\":4.43,\"end\":4.92}]"} {"utterance_id": "237-126133-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0000.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.45, 8.77, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":8.45,\"end\":8.77}]"} {"utterance_id": "5142-33396-0060", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0060.flac", "task_type": "lexical_detection", "query": "thorkel", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"thorkel\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.71, 1.14, "thorkel"]], "response_mode": "offline", "answer": "[{\"word\":\"thorkel\",\"start\":0.71,\"end\":1.14}]"} {"utterance_id": "4992-41806-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0017.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.09, 1.61, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":1.09,\"end\":1.61}]"} {"utterance_id": "1580-141083-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0021.flac", "task_type": "lexical_detection", "query": "guide", "prompt": "Once the audio has finished, identify every instance of the word \"guide\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.82, 3.21, "guide"]], "response_mode": "offline", "answer": "[{\"word\":\"guide\",\"start\":2.82,\"end\":3.21}]"} {"utterance_id": "672-122797-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0036.flac", "task_type": "lexical_detection", "query": "humpy", "prompt": "After listening to the whole recording, output every instance of the word \"humpy\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 0.97, "humpy"]], "response_mode": "offline", "answer": "[{\"word\":\"humpy\",\"start\":0.52,\"end\":0.97}]"} {"utterance_id": "8455-210777-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0051.flac", "task_type": "lexical_detection", "query": "speech", "prompt": "After the audio ends, extract all occurrences of the word \"speech\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.11, 2.55, "speech"]], "response_mode": "offline", "answer": "[{\"word\":\"speech\",\"start\":2.11,\"end\":2.55}]"} {"utterance_id": "5639-40744-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0037.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.59, 1.94, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":1.59,\"end\":1.94}]"} {"utterance_id": "1580-141084-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0048.flac", "task_type": "lexical_detection", "query": "locked", "prompt": "Process the full audio offline. Find every instance of the word \"locked\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.46, 7.74, "locked"]], "response_mode": "offline", "answer": "[{\"word\":\"locked\",\"start\":7.46,\"end\":7.74}]"} {"utterance_id": "1089-134691-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0000.flac", "task_type": "lexical_detection", "query": "wait", "prompt": "Once the audio has finished, identify every instance of the word \"wait\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.82, 1.1, "wait"]], "response_mode": "offline", "answer": "[{\"word\":\"wait\",\"start\":0.82,\"end\":1.1}]"} {"utterance_id": "260-123286-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0018.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "After listening to the whole recording, output every instance of an animal name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.42, 3.95, "creatures"]], "response_mode": "offline", "answer": "[{\"word\":\"creatures\",\"start\":3.42,\"end\":3.95}]"} {"utterance_id": "61-70968-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0033.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.53, 2.89, "robin"]], "response_mode": "offline", "answer": "[{\"word\":\"robin\",\"start\":2.53,\"end\":2.89}]"} {"utterance_id": "4446-2271-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0011.flac", "task_type": "lexical_detection", "query": "engineer", "prompt": "Process the full audio offline. Find every instance of the word \"engineer\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.01, 3.59, "engineer"]], "response_mode": "offline", "answer": "[{\"word\":\"engineer\",\"start\":3.01,\"end\":3.59}]"} {"utterance_id": "7176-92135-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0038.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.1, 3.57, "actors"]], "response_mode": "offline", "answer": "[{\"word\":\"actors\",\"start\":3.1,\"end\":3.57}]"} {"utterance_id": "8555-284449-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0000.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Once the audio has finished, identify every instance of a color name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[9.3, 9.57, "blue"]], "response_mode": "offline", "answer": "[{\"word\":\"blue\",\"start\":9.3,\"end\":9.57}]"} {"utterance_id": "61-70968-0055", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0055.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.31, 0.66, "robin"]], "response_mode": "offline", "answer": "[{\"word\":\"robin\",\"start\":0.31,\"end\":0.66}]"} {"utterance_id": "3729-6852-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0030.flac", "task_type": "lexical_detection", "query": "anywhere", "prompt": "After the audio ends, extract all occurrences of the word \"anywhere\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.66, 2.04, "anywhere"]], "response_mode": "offline", "answer": "[{\"word\":\"anywhere\",\"start\":1.66,\"end\":2.04}]"} {"utterance_id": "2094-142345-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0003.flac", "task_type": "lexical_detection", "query": "bare", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"bare\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.21, 4.6, "bare"]], "response_mode": "offline", "answer": "[{\"word\":\"bare\",\"start\":4.21,\"end\":4.6}]"} {"utterance_id": "3575-170457-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0053.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "When the utterance is complete, report how a person's name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.65, 0.9, "one"], [3.2, 3.6, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.65,\"end\":0.9},{\"word\":\"two\",\"start\":3.2,\"end\":3.6}]"} {"utterance_id": "61-70970-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0027.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.38, 0.79, "robin"]], "response_mode": "offline", "answer": "[{\"word\":\"robin\",\"start\":0.38,\"end\":0.79}]"} {"utterance_id": "5683-32866-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0013.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After the audio ends, extract all occurrences of a color name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.64, 8.89, "pale"]], "response_mode": "offline", "answer": "[{\"word\":\"pale\",\"start\":8.64,\"end\":8.89}]"} {"utterance_id": "1580-141083-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0002.flac", "task_type": "lexical_detection", "query": "friend's", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"friend's\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.64, 0.96, "friend's"]], "response_mode": "offline", "answer": "[{\"word\":\"friend's\",\"start\":0.64,\"end\":0.96}]"} {"utterance_id": "7176-92135-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0000.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a profession with word/start/end. If none, reply [].", "labels": [[7.26, 8.23, "one"], [9.27, 9.65, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.26,\"end\":8.23},{\"word\":\"two\",\"start\":9.27,\"end\":9.65}]"} {"utterance_id": "7729-102255-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0044.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.33, 3.39, "captain martin"]], "response_mode": "offline", "answer": "[{\"word\":\"captain martin\",\"start\":2.33,\"end\":3.39}]"} {"utterance_id": "2094-142345-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0034.flac", "task_type": "lexical_detection", "query": "linen", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"linen\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.84, 1.18, "linen"]], "response_mode": "offline", "answer": "[{\"word\":\"linen\",\"start\":0.84,\"end\":1.18}]"} {"utterance_id": "2094-142345-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0038.flac", "task_type": "lexical_counting", "query": "running", "prompt": "When the utterance is complete, report how the word \"running\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.66, 7.14, "one"], [18.17, 18.46, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.66,\"end\":7.14},{\"word\":\"two\",\"start\":18.17,\"end\":18.46}]"} {"utterance_id": "8463-294825-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0015.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.95, 2.24, "eighth"]], "response_mode": "offline", "answer": "[{\"word\":\"eighth\",\"start\":1.95,\"end\":2.24}]"} {"utterance_id": "2094-142345-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0016.flac", "task_type": "lexical_detection", "query": "spinning", "prompt": "Once the audio has finished, identify every instance of the word \"spinning\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.51, 1.19, "spinning"]], "response_mode": "offline", "answer": "[{\"word\":\"spinning\",\"start\":0.51,\"end\":1.19}]"} {"utterance_id": "5639-40744-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0009.flac", "task_type": "lexical_detection", "query": "hear", "prompt": "When the utterance is complete, report all instances of the word \"hear\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.68, 1.89, "hear"]], "response_mode": "offline", "answer": "[{\"word\":\"hear\",\"start\":1.68,\"end\":1.89}]"} {"utterance_id": "237-134493-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0011.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the full audio. After it finishes, find every instance of a color name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 0.84, "brown"]], "response_mode": "offline", "answer": "[{\"word\":\"brown\",\"start\":0.45,\"end\":0.84}]"} {"utterance_id": "1221-135766-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.92, 10.34, "hester"], [13.07, 13.43, "pearl"]], "response_mode": "offline", "answer": "[{\"word\":\"hester\",\"start\":9.92,\"end\":10.34},{\"word\":\"pearl\",\"start\":13.07,\"end\":13.43}]"} {"utterance_id": "1089-134691-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0004.flac", "task_type": "lexical_detection", "query": "slow", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"slow\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.4, 4.07, "slow"]], "response_mode": "offline", "answer": "[{\"word\":\"slow\",\"start\":3.4,\"end\":4.07}]"} {"utterance_id": "7021-85628-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0026.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Once the audio has finished, identify every instance of a family relation. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.41, 1.79, "son"]], "response_mode": "offline", "answer": "[{\"word\":\"son\",\"start\":1.41,\"end\":1.79}]"} {"utterance_id": "6829-68769-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0038.flac", "task_type": "lexical_detection", "query": "man's", "prompt": "When the utterance is complete, report all instances of the word \"man's\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.02, 1.32, "man's"]], "response_mode": "offline", "answer": "[{\"word\":\"man's\",\"start\":1.02,\"end\":1.32}]"} {"utterance_id": "5639-40744-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0008.flac", "task_type": "lexical_detection", "query": "touches", "prompt": "Once the audio has finished, identify every instance of the word \"touches\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.41, 0.85, "touches"]], "response_mode": "offline", "answer": "[{\"word\":\"touches\",\"start\":0.41,\"end\":0.85}]"} {"utterance_id": "6930-76324-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0014.flac", "task_type": "lexical_detection", "query": "pair", "prompt": "When the utterance is complete, report all instances of the word \"pair\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.82, 4.26, "pair"]], "response_mode": "offline", "answer": "[{\"word\":\"pair\",\"start\":3.82,\"end\":4.26}]"} {"utterance_id": "8463-294828-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0022.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.06, 2.18, "abraham lincoln"]], "response_mode": "offline", "answer": "[{\"word\":\"abraham lincoln\",\"start\":1.06,\"end\":2.18}]"} {"utterance_id": "1089-134686-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0015.flac", "task_type": "lexical_detection", "query": "dusk", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"dusk\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.38, 0.83, "dusk"]], "response_mode": "offline", "answer": "[{\"word\":\"dusk\",\"start\":0.38,\"end\":0.83}]"} {"utterance_id": "4970-29093-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0009.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an organization name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.6, 3.63, "astor library"]], "response_mode": "offline", "answer": "[{\"word\":\"astor library\",\"start\":2.6,\"end\":3.63}]"} {"utterance_id": "237-134500-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0029.flac", "task_type": "lexical_detection", "query": "around", "prompt": "Process the full audio offline. Find every instance of the word \"around\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.01, 1.28, "around"]], "response_mode": "offline", "answer": "[{\"word\":\"around\",\"start\":1.01,\"end\":1.28}]"} {"utterance_id": "121-127105-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0034.flac", "task_type": "lexical_detection", "query": "sounded", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"sounded\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.53, 0.92, "sounded"], [1.94, 2.32, "sounded"]], "response_mode": "offline", "answer": "[{\"word\":\"sounded\",\"start\":0.53,\"end\":0.92},{\"word\":\"sounded\",\"start\":1.94,\"end\":2.32}]"} {"utterance_id": "4970-29095-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0008.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "You will hear a complete utterance. When it ends, list every instance of a family relation you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.34, 0.82, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":0.34,\"end\":0.82}]"} {"utterance_id": "5683-32865-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0017.flac", "task_type": "lexical_detection", "query": "following", "prompt": "After the audio ends, extract all occurrences of the word \"following\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.96, 4.55, "following"]], "response_mode": "offline", "answer": "[{\"word\":\"following\",\"start\":3.96,\"end\":4.55}]"} {"utterance_id": "7176-92135-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0025.flac", "task_type": "lexical_counting", "query": "animal", "prompt": "When the utterance is complete, report how the word \"animal\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.66, 8.19, "one"], [11.05, 11.47, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.66,\"end\":8.19},{\"word\":\"two\",\"start\":11.05,\"end\":11.47}]"} {"utterance_id": "2830-3980-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0004.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "When the utterance is complete, report all instances of an organization name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.67, 1.96, "church"]], "response_mode": "offline", "answer": "[{\"word\":\"church\",\"start\":1.67,\"end\":1.96}]"} {"utterance_id": "1089-134686-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0031.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.48, 1.02, "stephen's"]], "response_mode": "offline", "answer": "[{\"word\":\"stephen's\",\"start\":0.48,\"end\":1.02}]"} {"utterance_id": "2830-3979-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0006.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.45, 2.84, "luther's"]], "response_mode": "offline", "answer": "[{\"word\":\"luther's\",\"start\":2.45,\"end\":2.84}]"} {"utterance_id": "8230-279154-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0004.flac", "task_type": "lexical_detection", "query": "population", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"population\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[9.59, 10.46, "population"]], "response_mode": "offline", "answer": "[{\"word\":\"population\",\"start\":9.59,\"end\":10.46}]"} {"utterance_id": "5105-28240-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0001.flac", "task_type": "lexical_detection", "query": "vessel", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"vessel\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.47, 1.87, "vessel"]], "response_mode": "offline", "answer": "[{\"word\":\"vessel\",\"start\":1.47,\"end\":1.87}]"} {"utterance_id": "260-123440-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0010.flac", "task_type": "lexical_detection", "query": "cheerfully", "prompt": "After the audio ends, extract all occurrences of the word \"cheerfully\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.41, 1.22, "cheerfully"]], "response_mode": "offline", "answer": "[{\"word\":\"cheerfully\",\"start\":0.41,\"end\":1.22}]"} {"utterance_id": "1580-141084-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0001.flac", "task_type": "lexical_detection", "query": "swiftly", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"swiftly\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.02, 1.61, "swiftly"]], "response_mode": "offline", "answer": "[{\"word\":\"swiftly\",\"start\":1.02,\"end\":1.61}]"} {"utterance_id": "4446-2273-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0021.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a plant name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.93, 2.57, "flowers"]], "response_mode": "offline", "answer": "[{\"word\":\"flowers\",\"start\":1.93,\"end\":2.57}]"} {"utterance_id": "5142-33396-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0033.flac", "task_type": "lexical_detection", "query": "ale", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"ale\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.1, 3.46, "ale"]], "response_mode": "offline", "answer": "[{\"word\":\"ale\",\"start\":3.1,\"end\":3.46}]"} {"utterance_id": "61-70968-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.15, 4.92, "master will"]], "response_mode": "offline", "answer": "[{\"word\":\"master will\",\"start\":4.15,\"end\":4.92}]"} {"utterance_id": "3575-170457-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0028.flac", "task_type": "semantic_detection", "query": "city", "prompt": "After the audio ends, extract all occurrences of a city name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 1.15, "keswick"]], "response_mode": "offline", "answer": "[{\"word\":\"keswick\",\"start\":0.51,\"end\":1.15}]"} {"utterance_id": "8224-274381-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0013.flac", "task_type": "lexical_counting", "query": "new", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"new\" (word, start, end). If none, reply [].", "labels": [[1.57, 1.87, "one"], [2.87, 3.15, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.57,\"end\":1.87},{\"word\":\"two\",\"start\":2.87,\"end\":3.15}]"} {"utterance_id": "4970-29095-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0001.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.61, 13.0, "philip's"]], "response_mode": "offline", "answer": "[{\"word\":\"philip's\",\"start\":12.61,\"end\":13.0}]"} {"utterance_id": "1221-135767-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0022.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.5, 1.17, "proprietor"]], "response_mode": "offline", "answer": "[{\"word\":\"proprietor\",\"start\":0.5,\"end\":1.17}]"} {"utterance_id": "4507-16021-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0022.flac", "task_type": "lexical_detection", "query": "lady", "prompt": "When the utterance is complete, report all instances of the word \"lady\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.79, 3.3, "lady"]], "response_mode": "offline", "answer": "[{\"word\":\"lady\",\"start\":2.79,\"end\":3.3}]"} {"utterance_id": "8463-294828-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0007.flac", "task_type": "lexical_counting", "query": "versed", "prompt": "After listening to the whole recording, output timed count words for the word \"versed\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[4.66, 5.01, "one"], [7.67, 8.07, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.66,\"end\":5.01},{\"word\":\"two\",\"start\":7.67,\"end\":8.07}]"} {"utterance_id": "1188-133604-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0035.flac", "task_type": "lexical_detection", "query": "dream", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"dream\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.87, 2.41, "dream"]], "response_mode": "offline", "answer": "[{\"word\":\"dream\",\"start\":1.87,\"end\":2.41}]"} {"utterance_id": "8230-279154-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0029.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Do not answer until the audio ends. Then list each instance of an ordinal number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.71, 5.35, "previous"]], "response_mode": "offline", "answer": "[{\"word\":\"previous\",\"start\":4.71,\"end\":5.35}]"} {"utterance_id": "260-123288-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0017.flac", "task_type": "lexical_detection", "query": "final", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"final\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.25, 4.86, "final"]], "response_mode": "offline", "answer": "[{\"word\":\"final\",\"start\":4.25,\"end\":4.86}]"} {"utterance_id": "1089-134686-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0027.flac", "task_type": "lexical_detection", "query": "clasped", "prompt": "Once the audio has finished, identify every instance of the word \"clasped\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.38, 0.91, "clasped"]], "response_mode": "offline", "answer": "[{\"word\":\"clasped\",\"start\":0.38,\"end\":0.91}]"} {"utterance_id": "4970-29093-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0008.flac", "task_type": "lexical_detection", "query": "ladder", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"ladder\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.6, 3.09, "ladder"]], "response_mode": "offline", "answer": "[{\"word\":\"ladder\",\"start\":2.6,\"end\":3.09}]"} {"utterance_id": "4446-2275-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0033.flac", "task_type": "lexical_counting", "query": "matter", "prompt": "Once the audio has finished, list each occurrence of the word \"matter\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[4.01, 4.35, "one"], [5.53, 5.8, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.01,\"end\":4.35},{\"word\":\"two\",\"start\":5.53,\"end\":5.8}]"} {"utterance_id": "672-122797-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0028.flac", "task_type": "lexical_detection", "query": "evening", "prompt": "After listening to the whole recording, output every instance of the word \"evening\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.93, 1.54, "evening"]], "response_mode": "offline", "answer": "[{\"word\":\"evening\",\"start\":0.93,\"end\":1.54}]"} {"utterance_id": "237-126133-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0017.flac", "task_type": "lexical_detection", "query": "screamed", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"screamed\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.73, 4.24, "screamed"]], "response_mode": "offline", "answer": "[{\"word\":\"screamed\",\"start\":3.73,\"end\":4.24}]"} {"utterance_id": "6829-68769-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0030.flac", "task_type": "lexical_detection", "query": "blank", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"blank\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.63, 2.94, "blank"]], "response_mode": "offline", "answer": "[{\"word\":\"blank\",\"start\":2.63,\"end\":2.94}]"} {"utterance_id": "3575-170457-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0021.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the full audio. After it finishes, find every instance of an organization name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.32, 2.71, "school"]], "response_mode": "offline", "answer": "[{\"word\":\"school\",\"start\":2.32,\"end\":2.71}]"} {"utterance_id": "260-123286-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0018.flac", "task_type": "semantic_detection", "query": "city", "prompt": "After listening to the whole recording, output every instance of a city name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.3, 1.72, "hamburg"]], "response_mode": "offline", "answer": "[{\"word\":\"hamburg\",\"start\":1.3,\"end\":1.72}]"} {"utterance_id": "7729-102255-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0023.flac", "task_type": "lexical_detection", "query": "characters", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"characters\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.03, 1.71, "characters"]], "response_mode": "offline", "answer": "[{\"word\":\"characters\",\"start\":1.03,\"end\":1.71}]"} {"utterance_id": "2830-3980-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0032.flac", "task_type": "lexical_counting", "query": "hath", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"hath\" (word, start, end). If none, reply [].", "labels": [[2.03, 2.29, "one"], [5.04, 5.28, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.03,\"end\":2.29},{\"word\":\"two\",\"start\":5.04,\"end\":5.28}]"} {"utterance_id": "8555-284447-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0015.flac", "task_type": "lexical_counting", "query": "both", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"both\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[11.33, 11.73, "one"], [14.24, 14.6, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":11.33,\"end\":11.73},{\"word\":\"two\",\"start\":14.24,\"end\":14.6}]"} {"utterance_id": "3729-6852-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0000.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.04, 2.58, "son"], [6.47, 7.09, "relatives"]], "response_mode": "offline", "answer": "[{\"word\":\"son\",\"start\":2.04,\"end\":2.58},{\"word\":\"relatives\",\"start\":6.47,\"end\":7.09}]"} {"utterance_id": "7176-92135-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0020.flac", "task_type": "lexical_detection", "query": "hallo", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"hallo\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.93, 5.29, "hallo"]], "response_mode": "offline", "answer": "[{\"word\":\"hallo\",\"start\":4.93,\"end\":5.29}]"} {"utterance_id": "4077-13751-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0006.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.28, 1.75, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":1.28,\"end\":1.75}]"} {"utterance_id": "7176-92135-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0026.flac", "task_type": "lexical_detection", "query": "hamlet", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"hamlet\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.76, 1.24, "hamlet"]], "response_mode": "offline", "answer": "[{\"word\":\"hamlet\",\"start\":0.76,\"end\":1.24}]"} {"utterance_id": "8555-284447-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0021.flac", "task_type": "lexical_detection", "query": "everyone", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"everyone\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[10.95, 11.43, "everyone"]], "response_mode": "offline", "answer": "[{\"word\":\"everyone\",\"start\":10.95,\"end\":11.43}]"} {"utterance_id": "8555-292519-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0004.flac", "task_type": "lexical_detection", "query": "pity", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"pity\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.68, 1.15, "pity"]], "response_mode": "offline", "answer": "[{\"word\":\"pity\",\"start\":0.68,\"end\":1.15}]"} {"utterance_id": "6930-75918-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0003.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.87, 4.35, "queen"], [11.8, 12.2, "queen"]], "response_mode": "offline", "answer": "[{\"word\":\"queen\",\"start\":3.87,\"end\":4.35},{\"word\":\"queen\",\"start\":11.8,\"end\":12.2}]"} {"utterance_id": "672-122797-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0044.flac", "task_type": "lexical_detection", "query": "against", "prompt": "After the audio ends, extract all occurrences of the word \"against\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.18, 1.58, "against"]], "response_mode": "offline", "answer": "[{\"word\":\"against\",\"start\":1.18,\"end\":1.58}]"} {"utterance_id": "1089-134686-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0019.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen carefully to the entire clip, then report every instance of an ordinal number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.11, 8.58, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":8.11,\"end\":8.58}]"} {"utterance_id": "5142-33396-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0059.flac", "task_type": "lexical_detection", "query": "frowned", "prompt": "After the audio ends, extract all occurrences of the word \"frowned\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.22, 4.53, "frowned"]], "response_mode": "offline", "answer": "[{\"word\":\"frowned\",\"start\":4.22,\"end\":4.53}]"} {"utterance_id": "1580-141084-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0029.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Once the audio has finished, identify every instance of a color name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.94, 1.22, "blue"]], "response_mode": "offline", "answer": "[{\"word\":\"blue\",\"start\":0.94,\"end\":1.22}]"} {"utterance_id": "2094-142345-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0012.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.17, 3.52, "niece"]], "response_mode": "offline", "answer": "[{\"word\":\"niece\",\"start\":3.17,\"end\":3.52}]"} {"utterance_id": "3729-6852-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0045.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.57, 14.9, "mouth"]], "response_mode": "offline", "answer": "[{\"word\":\"mouth\",\"start\":14.57,\"end\":14.9}]"} {"utterance_id": "1580-141084-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0014.flac", "task_type": "lexical_detection", "query": "bannister", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"bannister\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.56, 1.22, "bannister"]], "response_mode": "offline", "answer": "[{\"word\":\"bannister\",\"start\":0.56,\"end\":1.22}]"} {"utterance_id": "4992-41797-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0008.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, list every instance of an organization name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.09, 8.6, "lords"]], "response_mode": "offline", "answer": "[{\"word\":\"lords\",\"start\":8.09,\"end\":8.6}]"} {"utterance_id": "672-122797-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0012.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the full audio. After it finishes, find every instance of a plant name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.13, 6.59, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":6.13,\"end\":6.59}]"} {"utterance_id": "1284-1180-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0006.flac", "task_type": "lexical_detection", "query": "munchkins", "prompt": "When the utterance is complete, report all instances of the word \"munchkins\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.95, 3.64, "munchkins"]], "response_mode": "offline", "answer": "[{\"word\":\"munchkins\",\"start\":2.95,\"end\":3.64}]"} {"utterance_id": "8455-210777-0067", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0067.flac", "task_type": "lexical_detection", "query": "habits", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"habits\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.72, 1.1, "habits"]], "response_mode": "offline", "answer": "[{\"word\":\"habits\",\"start\":0.72,\"end\":1.1}]"} {"utterance_id": "61-70970-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0034.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.82, 3.29, "warrenton"]], "response_mode": "offline", "answer": "[{\"word\":\"warrenton\",\"start\":2.82,\"end\":3.29}]"} {"utterance_id": "1320-122617-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0011.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen to the full audio. After it finishes, find every instance of a city name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.58, 4.06, "village"]], "response_mode": "offline", "answer": "[{\"word\":\"village\",\"start\":3.58,\"end\":4.06}]"} {"utterance_id": "6829-68771-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0023.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.72, 3.04, "beth"]], "response_mode": "offline", "answer": "[{\"word\":\"beth\",\"start\":2.72,\"end\":3.04}]"} {"utterance_id": "1320-122617-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0020.flac", "task_type": "lexical_detection", "query": "prudent", "prompt": "Once the audio has finished, identify every instance of the word \"prudent\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.87, 5.36, "prudent"]], "response_mode": "offline", "answer": "[{\"word\":\"prudent\",\"start\":4.87,\"end\":5.36}]"} {"utterance_id": "237-134500-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0006.flac", "task_type": "lexical_detection", "query": "roses", "prompt": "Once the audio has finished, identify every instance of the word \"roses\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.16, 1.62, "roses"]], "response_mode": "offline", "answer": "[{\"word\":\"roses\",\"start\":1.16,\"end\":1.62}]"} {"utterance_id": "1284-1180-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0001.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the full audio. After it finishes, find every instance of a color name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.06, 5.54, "golden"]], "response_mode": "offline", "answer": "[{\"word\":\"golden\",\"start\":5.06,\"end\":5.54}]"} {"utterance_id": "908-31957-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0021.flac", "task_type": "lexical_detection", "query": "future", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"future\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.09, 3.57, "future"]], "response_mode": "offline", "answer": "[{\"word\":\"future\",\"start\":3.09,\"end\":3.57}]"} {"utterance_id": "2961-960-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.13, 3.74, "plato"]], "response_mode": "offline", "answer": "[{\"word\":\"plato\",\"start\":3.13,\"end\":3.74}]"} {"utterance_id": "1320-122617-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0014.flac", "task_type": "lexical_detection", "query": "motioned", "prompt": "After the audio ends, extract all occurrences of the word \"motioned\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.41, 2.85, "motioned"]], "response_mode": "offline", "answer": "[{\"word\":\"motioned\",\"start\":2.41,\"end\":2.85}]"} {"utterance_id": "4970-29093-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0021.flac", "task_type": "lexical_detection", "query": "nearer", "prompt": "After the audio ends, extract all occurrences of the word \"nearer\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.19, 1.52, "nearer"]], "response_mode": "offline", "answer": "[{\"word\":\"nearer\",\"start\":1.19,\"end\":1.52}]"} {"utterance_id": "121-121726-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0008.flac", "task_type": "lexical_detection", "query": "hose", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"hose\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 1.24, "hose"]], "response_mode": "offline", "answer": "[{\"word\":\"hose\",\"start\":0.52,\"end\":1.24}]"} {"utterance_id": "7176-88083-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0021.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[10.01, 10.37, "head"]], "response_mode": "offline", "answer": "[{\"word\":\"head\",\"start\":10.01,\"end\":10.37}]"} {"utterance_id": "1995-1836-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0004.flac", "task_type": "lexical_detection", "query": "tonight", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"tonight\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[7.27, 7.74, "tonight"]], "response_mode": "offline", "answer": "[{\"word\":\"tonight\",\"start\":7.27,\"end\":7.74}]"} {"utterance_id": "1995-1837-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.19, 2.71, "zora's"]], "response_mode": "offline", "answer": "[{\"word\":\"zora's\",\"start\":2.19,\"end\":2.71}]"} {"utterance_id": "6829-68769-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0053.flac", "task_type": "semantic_counting", "query": "money", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a money or currency term: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[4.1, 4.49, "one"], [5.58, 6.15, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.1,\"end\":4.49},{\"word\":\"two\",\"start\":5.58,\"end\":6.15}]"} {"utterance_id": "8224-274384-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0004.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "When the utterance is complete, report all instances of an ordinal number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[16.73, 17.22, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":16.73,\"end\":17.22}]"} {"utterance_id": "237-134500-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0036.flac", "task_type": "lexical_detection", "query": "marie", "prompt": "After the audio ends, extract all occurrences of the word \"marie\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.0, 6.38, "marie"]], "response_mode": "offline", "answer": "[{\"word\":\"marie\",\"start\":6.0,\"end\":6.38}]"} {"utterance_id": "61-70970-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0033.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen to the full audio. After it finishes, find every instance of an animal name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.97, 1.28, "horse"]], "response_mode": "offline", "answer": "[{\"word\":\"horse\",\"start\":0.97,\"end\":1.28}]"} {"utterance_id": "8463-294828-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0027.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "After listening to the whole recording, output every instance of an animal name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.12, 3.63, "animals"]], "response_mode": "offline", "answer": "[{\"word\":\"animals\",\"start\":3.12,\"end\":3.63}]"} {"utterance_id": "4507-16021-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0003.flac", "task_type": "semantic_counting", "query": "kinship", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a family relation with word/start/end. If none, reply [].", "labels": [[1.06, 1.59, "one"], [3.22, 3.79, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.06,\"end\":1.59},{\"word\":\"two\",\"start\":3.22,\"end\":3.79}]"} {"utterance_id": "4992-41797-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0003.flac", "task_type": "lexical_detection", "query": "popham", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"popham\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.94, 1.42, "popham"]], "response_mode": "offline", "answer": "[{\"word\":\"popham\",\"start\":0.94,\"end\":1.42}]"} {"utterance_id": "237-134500-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0003.flac", "task_type": "lexical_detection", "query": "sparkling", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"sparkling\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.03, 1.64, "sparkling"]], "response_mode": "offline", "answer": "[{\"word\":\"sparkling\",\"start\":1.03,\"end\":1.64}]"} {"utterance_id": "237-134493-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0017.flac", "task_type": "lexical_detection", "query": "divide", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"divide\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.47, 5.09, "divide"]], "response_mode": "offline", "answer": "[{\"word\":\"divide\",\"start\":4.47,\"end\":5.09}]"} {"utterance_id": "3575-170457-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0040.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a family relation with fields word, start, end (seconds). If none, reply [].", "labels": [[4.27, 4.75, "sisters"]], "response_mode": "offline", "answer": "[{\"word\":\"sisters\",\"start\":4.27,\"end\":4.75}]"} {"utterance_id": "6829-68769-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0035.flac", "task_type": "lexical_detection", "query": "find", "prompt": "After listening to the whole recording, output every instance of the word \"find\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.77, 2.07, "find"]], "response_mode": "offline", "answer": "[{\"word\":\"find\",\"start\":1.77,\"end\":2.07}]"} {"utterance_id": "3729-6852-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0001.flac", "task_type": "lexical_detection", "query": "illustrious", "prompt": "Once the audio has finished, identify every instance of the word \"illustrious\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.52, 6.18, "illustrious"]], "response_mode": "offline", "answer": "[{\"word\":\"illustrious\",\"start\":5.52,\"end\":6.18}]"} {"utterance_id": "1284-134647-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0004.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.15, 9.44, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":9.15,\"end\":9.44}]"} {"utterance_id": "1320-122617-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0006.flac", "task_type": "lexical_detection", "query": "freely", "prompt": "After listening to the whole recording, output every instance of the word \"freely\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.96, 3.45, "freely"]], "response_mode": "offline", "answer": "[{\"word\":\"freely\",\"start\":2.96,\"end\":3.45}]"} {"utterance_id": "121-123852-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123852/121-123852-0003.flac", "task_type": "lexical_counting", "query": "thought", "prompt": "Listen to the entire clip, then report the count sequence for the word \"thought\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.15, 0.62, "one"], [2.01, 2.43, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.15,\"end\":0.62},{\"word\":\"two\",\"start\":2.01,\"end\":2.43}]"} {"utterance_id": "2300-131720-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0027.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.25, 0.7, "edison"]], "response_mode": "offline", "answer": "[{\"word\":\"edison\",\"start\":0.25,\"end\":0.7}]"} {"utterance_id": "1995-1836-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0004.flac", "task_type": "lexical_counting", "query": "creighton", "prompt": "Listen to the entire clip, then report the count sequence for the word \"creighton\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.04, 11.56, "one"], [13.72, 14.23, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":11.04,\"end\":11.56},{\"word\":\"two\",\"start\":13.72,\"end\":14.23}]"} {"utterance_id": "237-134493-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0012.flac", "task_type": "lexical_detection", "query": "pick", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"pick\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.27, 2.51, "pick"]], "response_mode": "offline", "answer": "[{\"word\":\"pick\",\"start\":2.27,\"end\":2.51}]"} {"utterance_id": "237-126133-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0000.flac", "task_type": "lexical_detection", "query": "comforted", "prompt": "When the utterance is complete, report all instances of the word \"comforted\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.5, 2.03, "comforted"]], "response_mode": "offline", "answer": "[{\"word\":\"comforted\",\"start\":1.5,\"end\":2.03}]"} {"utterance_id": "7729-102255-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0044.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[6.71, 7.16, "jailer"]], "response_mode": "offline", "answer": "[{\"word\":\"jailer\",\"start\":6.71,\"end\":7.16}]"} {"utterance_id": "4992-41806-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0016.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a food name with fields word, start, end (seconds). If none, reply [].", "labels": [[11.85, 12.35, "meals"]], "response_mode": "offline", "answer": "[{\"word\":\"meals\",\"start\":11.85,\"end\":12.35}]"} {"utterance_id": "4970-29095-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0005.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.31, 2.81, "ruth"]], "response_mode": "offline", "answer": "[{\"word\":\"ruth\",\"start\":2.31,\"end\":2.81}]"} {"utterance_id": "4507-16021-0056", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0056.flac", "task_type": "lexical_detection", "query": "necessarily", "prompt": "After listening to the whole recording, output every instance of the word \"necessarily\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.79, 4.9, "necessarily"]], "response_mode": "offline", "answer": "[{\"word\":\"necessarily\",\"start\":3.79,\"end\":4.9}]"} {"utterance_id": "908-157963-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0011.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "Process the full audio offline. For every occurrence of an animal name, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.45, 7.08, "one"], [9.66, 10.26, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.45,\"end\":7.08},{\"word\":\"two\",\"start\":9.66,\"end\":10.26}]"} {"utterance_id": "2830-3980-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0029.flac", "task_type": "lexical_detection", "query": "write", "prompt": "When the utterance is complete, report all instances of the word \"write\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.98, 7.43, "write"]], "response_mode": "offline", "answer": "[{\"word\":\"write\",\"start\":6.98,\"end\":7.43}]"} {"utterance_id": "8455-210777-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0024.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.12, 6.36, "few"]], "response_mode": "offline", "answer": "[{\"word\":\"few\",\"start\":6.12,\"end\":6.36}]"} {"utterance_id": "1580-141084-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0036.flac", "task_type": "lexical_detection", "query": "nothing", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"nothing\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.64, 2.02, "nothing"]], "response_mode": "offline", "answer": "[{\"word\":\"nothing\",\"start\":1.64,\"end\":2.02}]"} {"utterance_id": "3575-170457-0049", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0049.flac", "task_type": "lexical_detection", "query": "communicated", "prompt": "Once the audio has finished, identify every instance of the word \"communicated\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.23, 1.78, "communicated"]], "response_mode": "offline", "answer": "[{\"word\":\"communicated\",\"start\":1.23,\"end\":1.78}]"} {"utterance_id": "3729-6852-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0029.flac", "task_type": "lexical_detection", "query": "weeks", "prompt": "After the audio ends, extract all occurrences of the word \"weeks\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.24, 3.67, "weeks"]], "response_mode": "offline", "answer": "[{\"word\":\"weeks\",\"start\":3.24,\"end\":3.67}]"} {"utterance_id": "7021-79730-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0000.flac", "task_type": "lexical_detection", "query": "modes", "prompt": "After the audio ends, extract all occurrences of the word \"modes\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.7, 1.18, "modes"]], "response_mode": "offline", "answer": "[{\"word\":\"modes\",\"start\":0.7,\"end\":1.18}]"} {"utterance_id": "7021-79740-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0009.flac", "task_type": "lexical_detection", "query": "dolls", "prompt": "Process the full audio offline. Find every instance of the word \"dolls\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.97, 2.61, "dolls"]], "response_mode": "offline", "answer": "[{\"word\":\"dolls\",\"start\":1.97,\"end\":2.61}]"} {"utterance_id": "3575-170457-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0043.flac", "task_type": "lexical_counting", "query": "decision", "prompt": "Listen to the entire clip, then report the count sequence for the word \"decision\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.0, 6.46, "one"], [10.42, 11.06, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.0,\"end\":6.46},{\"word\":\"two\",\"start\":10.42,\"end\":11.06}]"} {"utterance_id": "3729-6852-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0024.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.46, 1.72, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.46,\"end\":1.72}]"} {"utterance_id": "7127-75946-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0026.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.41, 5.09, "dancers"], [12.95, 13.42, "actors"]], "response_mode": "offline", "answer": "[{\"word\":\"dancers\",\"start\":4.41,\"end\":5.09},{\"word\":\"actors\",\"start\":12.95,\"end\":13.42}]"} {"utterance_id": "3575-170457-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0010.flac", "task_type": "lexical_detection", "query": "rare", "prompt": "After listening to the whole recording, output every instance of the word \"rare\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.95, 4.38, "rare"]], "response_mode": "offline", "answer": "[{\"word\":\"rare\",\"start\":3.95,\"end\":4.38}]"} {"utterance_id": "8463-294828-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0001.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.84, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":0.53,\"end\":0.84}]"} {"utterance_id": "1221-135767-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0016.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.82, 8.78, "governor bellingham"]], "response_mode": "offline", "answer": "[{\"word\":\"governor bellingham\",\"start\":7.82,\"end\":8.78}]"} {"utterance_id": "1284-1180-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0006.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a country name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.35, 2.77, "country"], [3.92, 4.29, "country"]], "response_mode": "offline", "answer": "[{\"word\":\"country\",\"start\":2.35,\"end\":2.77},{\"word\":\"country\",\"start\":3.92,\"end\":4.29}]"} {"utterance_id": "5683-32865-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0010.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.1, 4.54, "captain"]], "response_mode": "offline", "answer": "[{\"word\":\"captain\",\"start\":4.1,\"end\":4.54}]"} {"utterance_id": "3729-6852-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0032.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen to the full audio. After it finishes, find every instance of a city name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.73, 13.3, "paris"]], "response_mode": "offline", "answer": "[{\"word\":\"paris\",\"start\":12.73,\"end\":13.3}]"} {"utterance_id": "7176-92135-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0039.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.98, 1.44, "matthews"]], "response_mode": "offline", "answer": "[{\"word\":\"matthews\",\"start\":0.98,\"end\":1.44}]"} {"utterance_id": "7127-75947-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0029.flac", "task_type": "lexical_detection", "query": "girls", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"girls\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.68, 1.03, "girls"]], "response_mode": "offline", "answer": "[{\"word\":\"girls\",\"start\":0.68,\"end\":1.03}]"} {"utterance_id": "8463-294825-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0003.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.26, 0.64, "nemo"]], "response_mode": "offline", "answer": "[{\"word\":\"nemo\",\"start\":0.26,\"end\":0.64}]"} {"utterance_id": "1188-133604-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0007.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Once the audio has finished, identify every instance of a color name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.59, 0.94, "brown"], [4.35, 4.77, "black"], [7.23, 7.87, "dark gray"], [11.16, 11.68, "white"]], "response_mode": "offline", "answer": "[{\"word\":\"brown\",\"start\":0.59,\"end\":0.94},{\"word\":\"black\",\"start\":4.35,\"end\":4.77},{\"word\":\"dark gray\",\"start\":7.23,\"end\":7.87},{\"word\":\"white\",\"start\":11.16,\"end\":11.68}]"} {"utterance_id": "3575-170457-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0028.flac", "task_type": "lexical_detection", "query": "madam", "prompt": "Process the full audio offline. Find every instance of the word \"madam\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.54, 5.01, "madam"]], "response_mode": "offline", "answer": "[{\"word\":\"madam\",\"start\":4.54,\"end\":5.01}]"} {"utterance_id": "1320-122612-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0016.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.87, 1.22, "uncas"]], "response_mode": "offline", "answer": "[{\"word\":\"uncas\",\"start\":0.87,\"end\":1.22}]"} {"utterance_id": "61-70968-0055", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0055.flac", "task_type": "lexical_detection", "query": "devices", "prompt": "Once the audio has finished, identify every instance of the word \"devices\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.83, 3.47, "devices"]], "response_mode": "offline", "answer": "[{\"word\":\"devices\",\"start\":2.83,\"end\":3.47}]"} {"utterance_id": "121-127105-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0002.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.5, 0.67, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.5,\"end\":0.67}]"} {"utterance_id": "4992-41797-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.52, 3.92, "cyril"]], "response_mode": "offline", "answer": "[{\"word\":\"cyril\",\"start\":3.52,\"end\":3.92}]"} {"utterance_id": "6829-68769-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0009.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.68, 4.13, "jailer"]], "response_mode": "offline", "answer": "[{\"word\":\"jailer\",\"start\":3.68,\"end\":4.13}]"} {"utterance_id": "5639-40744-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0010.flac", "task_type": "lexical_detection", "query": "wrong", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"wrong\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.64, 2.94, "wrong"]], "response_mode": "offline", "answer": "[{\"word\":\"wrong\",\"start\":2.64,\"end\":2.94}]"} {"utterance_id": "3575-170457-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0009.flac", "task_type": "lexical_detection", "query": "january", "prompt": "Once the audio has finished, identify every instance of the word \"january\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.5, 1.13, "january"]], "response_mode": "offline", "answer": "[{\"word\":\"january\",\"start\":0.5,\"end\":1.13}]"} {"utterance_id": "1089-134686-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0028.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.91, 5.34, "saint francis xavier"]], "response_mode": "offline", "answer": "[{\"word\":\"saint francis xavier\",\"start\":3.91,\"end\":5.34}]"} {"utterance_id": "5142-33396-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0038.flac", "task_type": "lexical_detection", "query": "playing", "prompt": "When the utterance is complete, report all instances of the word \"playing\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.84, 3.34, "playing"]], "response_mode": "offline", "answer": "[{\"word\":\"playing\",\"start\":2.84,\"end\":3.34}]"} {"utterance_id": "5683-32879-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0015.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.21, 1.71, "rachel"]], "response_mode": "offline", "answer": "[{\"word\":\"rachel\",\"start\":1.21,\"end\":1.71}]"} {"utterance_id": "1284-1180-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0021.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "When the utterance is complete, report how an animal name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.49, 1.78, "one"], [7.18, 7.56, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.49,\"end\":1.78},{\"word\":\"two\",\"start\":7.18,\"end\":7.56}]"} {"utterance_id": "1089-134691-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0016.flac", "task_type": "lexical_counting", "query": "voyaging", "prompt": "Listen to the entire clip, then report the count sequence for the word \"voyaging\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.79, 1.31, "one"], [5.54, 6.07, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.79,\"end\":1.31},{\"word\":\"two\",\"start\":5.54,\"end\":6.07}]"} {"utterance_id": "2300-131720-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0016.flac", "task_type": "lexical_detection", "query": "marine", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"marine\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.55, 9.07, "marine"]], "response_mode": "offline", "answer": "[{\"word\":\"marine\",\"start\":8.55,\"end\":9.07}]"} {"utterance_id": "2830-3980-0068", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0068.flac", "task_type": "lexical_detection", "query": "greatness", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"greatness\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.64, 1.11, "greatness"]], "response_mode": "offline", "answer": "[{\"word\":\"greatness\",\"start\":0.64,\"end\":1.11}]"} {"utterance_id": "1320-122617-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0027.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.28, 3.64, "david"]], "response_mode": "offline", "answer": "[{\"word\":\"david\",\"start\":3.28,\"end\":3.64}]"} {"utterance_id": "260-123286-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0026.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.13, 0.4, "two"], [4.68, 4.95, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":0.13,\"end\":0.4},{\"word\":\"two\",\"start\":4.68,\"end\":4.95}]"} {"utterance_id": "3575-170457-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0027.flac", "task_type": "lexical_detection", "query": "gratification", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"gratification\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.87, 2.63, "gratification"]], "response_mode": "offline", "answer": "[{\"word\":\"gratification\",\"start\":1.87,\"end\":2.63}]"} {"utterance_id": "8455-210777-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0018.flac", "task_type": "lexical_detection", "query": "hospitality", "prompt": "Process the full audio offline. Find every instance of the word \"hospitality\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.76, 2.66, "hospitality"]], "response_mode": "offline", "answer": "[{\"word\":\"hospitality\",\"start\":1.76,\"end\":2.66}]"} {"utterance_id": "1089-134686-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0004.flac", "task_type": "lexical_detection", "query": "number", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"number\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.18, 0.65, "number"]], "response_mode": "offline", "answer": "[{\"word\":\"number\",\"start\":0.18,\"end\":0.65}]"} {"utterance_id": "1284-1180-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0012.flac", "task_type": "lexical_detection", "query": "lonelier", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"lonelier\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.36, 1.77, "lonelier"], [3.67, 4.21, "lonelier"]], "response_mode": "offline", "answer": "[{\"word\":\"lonelier\",\"start\":1.36,\"end\":1.77},{\"word\":\"lonelier\",\"start\":3.67,\"end\":4.21}]"} {"utterance_id": "5683-32865-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0010.flac", "task_type": "lexical_detection", "query": "ace", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"ace\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.61, "ace"]], "response_mode": "offline", "answer": "[{\"word\":\"ace\",\"start\":2.35,\"end\":2.61}]"} {"utterance_id": "61-70970-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0029.flac", "task_type": "lexical_detection", "query": "warrenton", "prompt": "After the audio ends, extract all occurrences of the word \"warrenton\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.95, 3.67, "warrenton"]], "response_mode": "offline", "answer": "[{\"word\":\"warrenton\",\"start\":2.95,\"end\":3.67}]"} {"utterance_id": "4446-2273-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0024.flac", "task_type": "lexical_detection", "query": "bell", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"bell\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.03, 2.22, "bell"]], "response_mode": "offline", "answer": "[{\"word\":\"bell\",\"start\":2.03,\"end\":2.22}]"} {"utterance_id": "8224-274381-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0011.flac", "task_type": "lexical_detection", "query": "conduct", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"conduct\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.77, 1.41, "conduct"]], "response_mode": "offline", "answer": "[{\"word\":\"conduct\",\"start\":0.77,\"end\":1.41}]"} {"utterance_id": "6930-76324-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.0, 2.48, "joyce"]], "response_mode": "offline", "answer": "[{\"word\":\"joyce\",\"start\":2.0,\"end\":2.48}]"} {"utterance_id": "8555-284447-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0021.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "When the utterance is complete, report all instances of an ordinal number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.48, 3.9, "final"]], "response_mode": "offline", "answer": "[{\"word\":\"final\",\"start\":3.48,\"end\":3.9}]"} {"utterance_id": "5142-36586-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36586/5142-36586-0000.flac", "task_type": "lexical_detection", "query": "manifest", "prompt": "Once the audio has finished, identify every instance of the word \"manifest\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.75, 1.36, "manifest"]], "response_mode": "offline", "answer": "[{\"word\":\"manifest\",\"start\":0.75,\"end\":1.36}]"} {"utterance_id": "8463-294825-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.6, 2.04, "nemo"]], "response_mode": "offline", "answer": "[{\"word\":\"nemo\",\"start\":1.6,\"end\":2.04}]"} {"utterance_id": "61-70968-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0001.flac", "task_type": "lexical_detection", "query": "earnest", "prompt": "Once the audio has finished, identify every instance of the word \"earnest\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.24, 1.66, "earnest"]], "response_mode": "offline", "answer": "[{\"word\":\"earnest\",\"start\":1.24,\"end\":1.66}]"} {"utterance_id": "5683-32879-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0011.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a body part: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[5.92, 6.28, "one"], [7.64, 8.05, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.92,\"end\":6.28},{\"word\":\"two\",\"start\":7.64,\"end\":8.05}]"} {"utterance_id": "1580-141083-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0011.flac", "task_type": "lexical_detection", "query": "tip", "prompt": "After listening to the whole recording, output every instance of the word \"tip\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.79, 1.01, "tip"]], "response_mode": "offline", "answer": "[{\"word\":\"tip\",\"start\":0.79,\"end\":1.01}]"} {"utterance_id": "4992-23283-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0020.flac", "task_type": "lexical_detection", "query": "vanquished", "prompt": "After listening to the whole recording, output every instance of the word \"vanquished\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.01, 2.54, "vanquished"]], "response_mode": "offline", "answer": "[{\"word\":\"vanquished\",\"start\":2.01,\"end\":2.54}]"} {"utterance_id": "6930-76324-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0028.flac", "task_type": "semantic_detection", "query": "date", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a date with fields word, start, end (seconds). If none, reply [].", "labels": [[4.86, 7.72, "april sixteenth eighteen sixty one"]], "response_mode": "offline", "answer": "[{\"word\":\"april sixteenth eighteen sixty one\",\"start\":4.86,\"end\":7.72}]"} {"utterance_id": "1188-133604-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0016.flac", "task_type": "lexical_detection", "query": "producing", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"producing\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.85, 14.32, "producing"]], "response_mode": "offline", "answer": "[{\"word\":\"producing\",\"start\":13.85,\"end\":14.32}]"} {"utterance_id": "7176-92135-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0042.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Once the audio has finished, identify every instance of a food name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.45, 2.77, "meals"]], "response_mode": "offline", "answer": "[{\"word\":\"meals\",\"start\":2.45,\"end\":2.77}]"} {"utterance_id": "7127-75946-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0022.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[3.69, 4.24, "servant"]], "response_mode": "offline", "answer": "[{\"word\":\"servant\",\"start\":3.69,\"end\":4.24}]"} {"utterance_id": "8463-294828-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0013.flac", "task_type": "lexical_detection", "query": "got", "prompt": "After listening to the whole recording, output every instance of the word \"got\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.0, 6.21, "got"]], "response_mode": "offline", "answer": "[{\"word\":\"got\",\"start\":6.0,\"end\":6.21}]"} {"utterance_id": "4992-23283-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0014.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.94, 2.7, "miss woodley"]], "response_mode": "offline", "answer": "[{\"word\":\"miss woodley\",\"start\":1.94,\"end\":2.7}]"} {"utterance_id": "6829-68769-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0017.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.76, "jailer"]], "response_mode": "offline", "answer": "[{\"word\":\"jailer\",\"start\":2.35,\"end\":2.76}]"} {"utterance_id": "7176-92135-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0000.flac", "task_type": "semantic_counting", "query": "money", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a money or currency term (word, start, end). If none, reply [].", "labels": [[11.38, 11.79, "one"], [13.1, 13.43, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":11.38,\"end\":11.79},{\"word\":\"two\",\"start\":13.1,\"end\":13.43}]"} {"utterance_id": "5639-40744-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0002.flac", "task_type": "lexical_detection", "query": "companions", "prompt": "After the audio ends, extract all occurrences of the word \"companions\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.14, 1.94, "companions"]], "response_mode": "offline", "answer": "[{\"word\":\"companions\",\"start\":1.14,\"end\":1.94}]"} {"utterance_id": "4970-29095-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.28, 2.51, "ruth"]], "response_mode": "offline", "answer": "[{\"word\":\"ruth\",\"start\":2.28,\"end\":2.51}]"} {"utterance_id": "5142-33396-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0028.flac", "task_type": "lexical_detection", "query": "bench", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"bench\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 0.8, "bench"]], "response_mode": "offline", "answer": "[{\"word\":\"bench\",\"start\":0.45,\"end\":0.8}]"} {"utterance_id": "3570-5694-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0018.flac", "task_type": "lexical_detection", "query": "liveries", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"liveries\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.65, 2.13, "liveries"]], "response_mode": "offline", "answer": "[{\"word\":\"liveries\",\"start\":1.65,\"end\":2.13}]"} {"utterance_id": "2961-960-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0002.flac", "task_type": "lexical_detection", "query": "depths", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"depths\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.44, 1.98, "depths"]], "response_mode": "offline", "answer": "[{\"word\":\"depths\",\"start\":1.44,\"end\":1.98}]"} {"utterance_id": "121-123859-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123859/121-123859-0000.flac", "task_type": "lexical_counting", "query": "none", "prompt": "After the audio ends, tally every instance of the word \"none\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[8.86, 9.23, "one"], [11.49, 11.76, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":8.86,\"end\":9.23},{\"word\":\"two\",\"start\":11.49,\"end\":11.76}]"} {"utterance_id": "5683-32879-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0021.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.39, 0.98, "dorcas"]], "response_mode": "offline", "answer": "[{\"word\":\"dorcas\",\"start\":0.39,\"end\":0.98}]"} {"utterance_id": "4507-16021-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0039.flac", "task_type": "lexical_detection", "query": "murmur", "prompt": "Once the audio has finished, identify every instance of the word \"murmur\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.45, 5.09, "murmur"]], "response_mode": "offline", "answer": "[{\"word\":\"murmur\",\"start\":4.45,\"end\":5.09}]"} {"utterance_id": "4992-23283-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0011.flac", "task_type": "lexical_detection", "query": "value", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"value\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.63, 3.18, "value"]], "response_mode": "offline", "answer": "[{\"word\":\"value\",\"start\":2.63,\"end\":3.18}]"} {"utterance_id": "7176-92135-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0045.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.42, 3.81, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":3.42,\"end\":3.81}]"} {"utterance_id": "6829-68771-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0019.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.9, 3.23, "maids"]], "response_mode": "offline", "answer": "[{\"word\":\"maids\",\"start\":2.9,\"end\":3.23}]"} {"utterance_id": "2961-960-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0011.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Process the full audio offline. For every occurrence of a person's name, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.69, 1.63, "one"], [5.01, 5.99, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.69,\"end\":1.63},{\"word\":\"two\",\"start\":5.01,\"end\":5.99}]"} {"utterance_id": "2094-142345-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0059.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.29, 3.85, "captain"]], "response_mode": "offline", "answer": "[{\"word\":\"captain\",\"start\":3.29,\"end\":3.85}]"} {"utterance_id": "4992-23283-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.6, 2.17, "miss milner's"]], "response_mode": "offline", "answer": "[{\"word\":\"miss milner's\",\"start\":1.6,\"end\":2.17}]"} {"utterance_id": "3575-170457-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0021.flac", "task_type": "lexical_detection", "query": "duty", "prompt": "Once the audio has finished, identify every instance of the word \"duty\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.52, 1.91, "duty"]], "response_mode": "offline", "answer": "[{\"word\":\"duty\",\"start\":1.52,\"end\":1.91}]"} {"utterance_id": "8230-279154-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0043.flac", "task_type": "lexical_counting", "query": "recognition", "prompt": "Once the audio has finished, list each occurrence of the word \"recognition\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[9.76, 10.59, "one"], [23.32, 24.35, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":9.76,\"end\":10.59},{\"word\":\"two\",\"start\":23.32,\"end\":24.35}]"} {"utterance_id": "4077-13751-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0020.flac", "task_type": "semantic_detection", "query": "date", "prompt": "You will hear a complete utterance. When it ends, list every instance of a date you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.81, 10.04, "march sixteenth eighteen thirty nine"]], "response_mode": "offline", "answer": "[{\"word\":\"march sixteenth eighteen thirty nine\",\"start\":7.81,\"end\":10.04}]"} {"utterance_id": "1188-133604-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0023.flac", "task_type": "semantic_counting", "query": "color", "prompt": "After listening to the whole recording, output timed count words for a color name (one, two, three, ...) as JSON. If none, reply [].", "labels": [[5.44, 6.0, "one"], [8.61, 9.16, "two"], [14.77, 15.24, "three"], [20.55, 20.84, "four"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.44,\"end\":6.0},{\"word\":\"two\",\"start\":8.61,\"end\":9.16},{\"word\":\"three\",\"start\":14.77,\"end\":15.24},{\"word\":\"four\",\"start\":20.55,\"end\":20.84}]"} {"utterance_id": "5105-28233-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28233/5105-28233-0005.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.81, 2.12, "foot"]], "response_mode": "offline", "answer": "[{\"word\":\"foot\",\"start\":1.81,\"end\":2.12}]"} {"utterance_id": "7021-85628-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0022.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.05, 4.51, "mouths"]], "response_mode": "offline", "answer": "[{\"word\":\"mouths\",\"start\":4.05,\"end\":4.51}]"} {"utterance_id": "5639-40744-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0025.flac", "task_type": "lexical_detection", "query": "spot", "prompt": "Process the full audio offline. Find every instance of the word \"spot\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.94, 8.4, "spot"]], "response_mode": "offline", "answer": "[{\"word\":\"spot\",\"start\":7.94,\"end\":8.4}]"} {"utterance_id": "3575-170457-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0052.flac", "task_type": "semantic_detection", "query": "date", "prompt": "After the audio ends, extract all occurrences of a date. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.11, 2.74, "christmas"]], "response_mode": "offline", "answer": "[{\"word\":\"christmas\",\"start\":2.11,\"end\":2.74}]"} {"utterance_id": "4970-29093-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0005.flac", "task_type": "lexical_detection", "query": "humbly", "prompt": "When the utterance is complete, report all instances of the word \"humbly\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.61, 5.05, "humbly"]], "response_mode": "offline", "answer": "[{\"word\":\"humbly\",\"start\":4.61,\"end\":5.05}]"} {"utterance_id": "4077-13751-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0007.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[10.3, 10.92, "missionaries"]], "response_mode": "offline", "answer": "[{\"word\":\"missionaries\",\"start\":10.3,\"end\":10.92}]"} {"utterance_id": "8455-210777-0049", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0049.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.04, 3.7, "officer"]], "response_mode": "offline", "answer": "[{\"word\":\"officer\",\"start\":3.04,\"end\":3.7}]"} {"utterance_id": "121-127105-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0019.flac", "task_type": "lexical_detection", "query": "expressed", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"expressed\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.21, 1.68, "expressed"]], "response_mode": "offline", "answer": "[{\"word\":\"expressed\",\"start\":1.21,\"end\":1.68}]"} {"utterance_id": "1188-133604-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0006.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.39, 1.66, "beak"]], "response_mode": "offline", "answer": "[{\"word\":\"beak\",\"start\":1.39,\"end\":1.66}]"} {"utterance_id": "672-122797-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0021.flac", "task_type": "semantic_detection", "query": "date", "prompt": "When the utterance is complete, report all instances of a date including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.85, 1.37, "christmas"]], "response_mode": "offline", "answer": "[{\"word\":\"christmas\",\"start\":0.85,\"end\":1.37}]"} {"utterance_id": "4507-16021-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0040.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Process the full audio offline. Find every instance of an animal name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.07, 2.84, "hydras"]], "response_mode": "offline", "answer": "[{\"word\":\"hydras\",\"start\":2.07,\"end\":2.84}]"} {"utterance_id": "6930-75918-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0007.flac", "task_type": "lexical_detection", "query": "always", "prompt": "After listening to the whole recording, output every instance of the word \"always\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.23, 2.52, "always"]], "response_mode": "offline", "answer": "[{\"word\":\"always\",\"start\":2.23,\"end\":2.52}]"} {"utterance_id": "5142-33396-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0050.flac", "task_type": "lexical_detection", "query": "ease", "prompt": "Once the audio has finished, identify every instance of the word \"ease\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.23, 1.51, "ease"]], "response_mode": "offline", "answer": "[{\"word\":\"ease\",\"start\":1.23,\"end\":1.51}]"} {"utterance_id": "5683-32866-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0019.flac", "task_type": "lexical_detection", "query": "survive", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"survive\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.33, 7.87, "survive"]], "response_mode": "offline", "answer": "[{\"word\":\"survive\",\"start\":7.33,\"end\":7.87}]"} {"utterance_id": "2830-3980-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0054.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a person's name with word/start/end. If none, reply [].", "labels": [[0.55, 0.94, "one"], [3.19, 3.46, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.55,\"end\":0.94},{\"word\":\"two\",\"start\":3.19,\"end\":3.46}]"} {"utterance_id": "7729-102255-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0011.flac", "task_type": "lexical_detection", "query": "movements", "prompt": "After listening to the whole recording, output every instance of the word \"movements\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.16, 3.72, "movements"]], "response_mode": "offline", "answer": "[{\"word\":\"movements\",\"start\":3.16,\"end\":3.72}]"} {"utterance_id": "4507-16021-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0032.flac", "task_type": "lexical_detection", "query": "severity", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"severity\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.36, 5.42, "severity"]], "response_mode": "offline", "answer": "[{\"word\":\"severity\",\"start\":4.36,\"end\":5.42}]"} {"utterance_id": "8455-210777-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0016.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.71, 3.29, "president"]], "response_mode": "offline", "answer": "[{\"word\":\"president\",\"start\":2.71,\"end\":3.29}]"} {"utterance_id": "4970-29093-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0014.flac", "task_type": "lexical_detection", "query": "going", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"going\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.28, 1.56, "going"]], "response_mode": "offline", "answer": "[{\"word\":\"going\",\"start\":1.28,\"end\":1.56}]"} {"utterance_id": "5639-40744-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0022.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.02, 7.34, "mother"], [10.8, 11.22, "son"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":7.02,\"end\":7.34},{\"word\":\"son\",\"start\":10.8,\"end\":11.22}]"} {"utterance_id": "908-157963-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0024.flac", "task_type": "lexical_detection", "query": "weakness", "prompt": "Once the audio has finished, identify every instance of the word \"weakness\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.69, 1.38, "weakness"]], "response_mode": "offline", "answer": "[{\"word\":\"weakness\",\"start\":0.69,\"end\":1.38}]"} {"utterance_id": "1221-135767-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0003.flac", "task_type": "lexical_detection", "query": "pig", "prompt": "After the audio ends, extract all occurrences of the word \"pig\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.74, 7.16, "pig"]], "response_mode": "offline", "answer": "[{\"word\":\"pig\",\"start\":6.74,\"end\":7.16}]"} {"utterance_id": "1188-133604-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0028.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a weather term with fields word, start, end (seconds). If none, reply [].", "labels": [[17.38, 17.84, "cloud"]], "response_mode": "offline", "answer": "[{\"word\":\"cloud\",\"start\":17.38,\"end\":17.84}]"} {"utterance_id": "8455-210777-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0043.flac", "task_type": "lexical_detection", "query": "mission", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"mission\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.47, 1.88, "mission"]], "response_mode": "offline", "answer": "[{\"word\":\"mission\",\"start\":1.47,\"end\":1.88}]"} {"utterance_id": "121-127105-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0031.flac", "task_type": "lexical_detection", "query": "vision", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"vision\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.04, 4.47, "vision"]], "response_mode": "offline", "answer": "[{\"word\":\"vision\",\"start\":4.04,\"end\":4.47}]"} {"utterance_id": "2830-3980-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.39, 0.65, "paul"]], "response_mode": "offline", "answer": "[{\"word\":\"paul\",\"start\":0.39,\"end\":0.65}]"} {"utterance_id": "4077-13751-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0011.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a city name with fields word, start, end (seconds). If none, reply [].", "labels": [[4.57, 5.29, "independence"]], "response_mode": "offline", "answer": "[{\"word\":\"independence\",\"start\":4.57,\"end\":5.29}]"} {"utterance_id": "1089-134686-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0004.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.65, 1.24, "ten"]], "response_mode": "offline", "answer": "[{\"word\":\"ten\",\"start\":0.65,\"end\":1.24}]"} {"utterance_id": "5142-33396-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0019.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After the audio ends, extract all occurrences of a plant name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.36, 3.82, "crops"]], "response_mode": "offline", "answer": "[{\"word\":\"crops\",\"start\":3.36,\"end\":3.82}]"} {"utterance_id": "8455-210777-0058", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0058.flac", "task_type": "lexical_detection", "query": "acknowledged", "prompt": "After listening to the whole recording, output every instance of the word \"acknowledged\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.49, 2.06, "acknowledged"]], "response_mode": "offline", "answer": "[{\"word\":\"acknowledged\",\"start\":1.49,\"end\":2.06}]"} {"utterance_id": "5683-32866-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0014.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.29, 1.75, "stanley"]], "response_mode": "offline", "answer": "[{\"word\":\"stanley\",\"start\":1.29,\"end\":1.75}]"} {"utterance_id": "1284-1181-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0010.flac", "task_type": "lexical_detection", "query": "leaving", "prompt": "Process the full audio offline. Find every instance of the word \"leaving\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.75, "leaving"]], "response_mode": "offline", "answer": "[{\"word\":\"leaving\",\"start\":2.35,\"end\":2.75}]"} {"utterance_id": "2830-3980-0049", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0049.flac", "task_type": "lexical_detection", "query": "embrace", "prompt": "Once the audio has finished, identify every instance of the word \"embrace\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.21, 0.77, "embrace"]], "response_mode": "offline", "answer": "[{\"word\":\"embrace\",\"start\":0.21,\"end\":0.77}]"} {"utterance_id": "7021-85628-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0019.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.56, 0.82, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.56,\"end\":0.82}]"} {"utterance_id": "1089-134686-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0012.flac", "task_type": "lexical_detection", "query": "mild", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"mild\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.43, 2.82, "mild"]], "response_mode": "offline", "answer": "[{\"word\":\"mild\",\"start\":2.43,\"end\":2.82}]"} {"utterance_id": "672-122797-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0025.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "Listen to the full audio. Count every occurrence of a plant name. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.61, 0.9, "one"], [4.95, 5.37, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.61,\"end\":0.9},{\"word\":\"two\",\"start\":4.95,\"end\":5.37}]"} {"utterance_id": "3729-6852-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0017.flac", "task_type": "lexical_detection", "query": "chairs", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"chairs\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.54, 1.98, "chairs"]], "response_mode": "offline", "answer": "[{\"word\":\"chairs\",\"start\":1.54,\"end\":1.98}]"} {"utterance_id": "4507-16021-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0047.flac", "task_type": "lexical_counting", "query": "day", "prompt": "After listening to the whole recording, output timed count words for the word \"day\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[4.93, 5.27, "one"], [10.83, 11.11, "two"], [14.9, 15.17, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.93,\"end\":5.27},{\"word\":\"two\",\"start\":10.83,\"end\":11.11},{\"word\":\"three\",\"start\":14.9,\"end\":15.17}]"} {"utterance_id": "2961-961-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0011.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.21, 6.86, "children's"]], "response_mode": "offline", "answer": "[{\"word\":\"children's\",\"start\":6.21,\"end\":6.86}]"} {"utterance_id": "5683-32879-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0016.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[9.19, 9.71, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":9.19,\"end\":9.71}]"} {"utterance_id": "8463-294825-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0002.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Do not answer until the audio ends. Then list each instance of a city name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.02, 1.49, "paris"]], "response_mode": "offline", "answer": "[{\"word\":\"paris\",\"start\":1.02,\"end\":1.49}]"} {"utterance_id": "1995-1837-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0027.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Do not answer until the audio ends. Then list each instance of a plant name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.19, 5.6, "cotton"]], "response_mode": "offline", "answer": "[{\"word\":\"cotton\",\"start\":5.19,\"end\":5.6}]"} {"utterance_id": "1188-133604-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0016.flac", "task_type": "lexical_counting", "query": "work", "prompt": "Process the full audio offline. For every occurrence of the word \"work\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.76, 3.1, "one"], [14.75, 15.07, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.76,\"end\":3.1},{\"word\":\"two\",\"start\":14.75,\"end\":15.07}]"} {"utterance_id": "7176-92135-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0002.flac", "task_type": "lexical_detection", "query": "moments", "prompt": "When the utterance is complete, report all instances of the word \"moments\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.8, 3.26, "moments"]], "response_mode": "offline", "answer": "[{\"word\":\"moments\",\"start\":2.8,\"end\":3.26}]"} {"utterance_id": "5142-36377-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0005.flac", "task_type": "lexical_detection", "query": "honestly", "prompt": "When the utterance is complete, report all instances of the word \"honestly\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.83, 4.23, "honestly"]], "response_mode": "offline", "answer": "[{\"word\":\"honestly\",\"start\":3.83,\"end\":4.23}]"} {"utterance_id": "7127-75947-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0032.flac", "task_type": "lexical_detection", "query": "perhaps", "prompt": "When the utterance is complete, report all instances of the word \"perhaps\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.45, 1.88, "perhaps"]], "response_mode": "offline", "answer": "[{\"word\":\"perhaps\",\"start\":1.45,\"end\":1.88}]"} {"utterance_id": "1221-135767-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0017.flac", "task_type": "lexical_detection", "query": "token", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"token\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.42, 1.81, "token"]], "response_mode": "offline", "answer": "[{\"word\":\"token\",\"start\":1.42,\"end\":1.81}]"} {"utterance_id": "260-123288-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0013.flac", "task_type": "semantic_counting", "query": "weather", "prompt": "Process the full audio offline. For every occurrence of a weather term, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.09, 8.61, "one"], [10.27, 11.03, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":8.09,\"end\":8.61},{\"word\":\"two\",\"start\":10.27,\"end\":11.03}]"} {"utterance_id": "5683-32879-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.91, 1.55, "rachel"]], "response_mode": "offline", "answer": "[{\"word\":\"rachel\",\"start\":0.91,\"end\":1.55}]"} {"utterance_id": "4077-13751-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0015.flac", "task_type": "lexical_detection", "query": "forty", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"forty\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.67, 10.18, "forty"]], "response_mode": "offline", "answer": "[{\"word\":\"forty\",\"start\":9.67,\"end\":10.18}]"} {"utterance_id": "1284-1180-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0030.flac", "task_type": "lexical_detection", "query": "mixture", "prompt": "Once the audio has finished, identify every instance of the word \"mixture\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[10.79, 11.17, "mixture"]], "response_mode": "offline", "answer": "[{\"word\":\"mixture\",\"start\":10.79,\"end\":11.17}]"} {"utterance_id": "7729-102255-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0014.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Process the full audio offline. Find every instance of a city name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.73, 5.16, "town"]], "response_mode": "offline", "answer": "[{\"word\":\"town\",\"start\":4.73,\"end\":5.16}]"} {"utterance_id": "5639-40744-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0030.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[9.99, 10.37, "wife"]], "response_mode": "offline", "answer": "[{\"word\":\"wife\",\"start\":9.99,\"end\":10.37}]"} {"utterance_id": "237-134500-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.33, 0.71, "frank"]], "response_mode": "offline", "answer": "[{\"word\":\"frank\",\"start\":0.33,\"end\":0.71}]"} {"utterance_id": "1580-141084-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0032.flac", "task_type": "lexical_detection", "query": "control", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"control\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.99, 3.53, "control"]], "response_mode": "offline", "answer": "[{\"word\":\"control\",\"start\":2.99,\"end\":3.53}]"} {"utterance_id": "5142-36586-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36586/5142-36586-0004.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.8, 3.34, "parts"]], "response_mode": "offline", "answer": "[{\"word\":\"parts\",\"start\":2.8,\"end\":3.34}]"} {"utterance_id": "1320-122617-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0039.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Once the audio has finished, identify every instance of an animal name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.84, 4.17, "animal"]], "response_mode": "offline", "answer": "[{\"word\":\"animal\",\"start\":3.84,\"end\":4.17}]"} {"utterance_id": "2094-142345-0057", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0057.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.17, 3.41, "ten"], [8.42, 8.64, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"ten\",\"start\":3.17,\"end\":3.41},{\"word\":\"one\",\"start\":8.42,\"end\":8.64}]"} {"utterance_id": "6829-68769-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0029.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Do not answer until the audio ends. Then list each instance of a money or currency term as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.52, 0.87, "stock"]], "response_mode": "offline", "answer": "[{\"word\":\"stock\",\"start\":0.52,\"end\":0.87}]"} {"utterance_id": "6930-75918-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0015.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.61, 3.02, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":2.61,\"end\":3.02}]"} {"utterance_id": "5142-33396-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0013.flac", "task_type": "lexical_detection", "query": "boy", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"boy\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.69, 1.18, "boy"]], "response_mode": "offline", "answer": "[{\"word\":\"boy\",\"start\":0.69,\"end\":1.18}]"} {"utterance_id": "237-126133-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0000.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.98, 7.54, "gardener"]], "response_mode": "offline", "answer": "[{\"word\":\"gardener\",\"start\":6.98,\"end\":7.54}]"} {"utterance_id": "1580-141083-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0004.flac", "task_type": "lexical_detection", "query": "text", "prompt": "When the utterance is complete, report all instances of the word \"text\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.21, 2.66, "text"]], "response_mode": "offline", "answer": "[{\"word\":\"text\",\"start\":2.21,\"end\":2.66}]"} {"utterance_id": "8463-294828-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0008.flac", "task_type": "lexical_detection", "query": "gallant", "prompt": "Once the audio has finished, identify every instance of the word \"gallant\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.49, 1.93, "gallant"]], "response_mode": "offline", "answer": "[{\"word\":\"gallant\",\"start\":1.49,\"end\":1.93}]"} {"utterance_id": "3729-6852-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0022.flac", "task_type": "lexical_detection", "query": "address", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"address\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.57, 0.89, "address"]], "response_mode": "offline", "answer": "[{\"word\":\"address\",\"start\":0.57,\"end\":0.89}]"} {"utterance_id": "260-123288-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0027.flac", "task_type": "lexical_counting", "query": "fills", "prompt": "Listen to the full audio. Count every occurrence of the word \"fills\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.27, 2.62, "one"], [5.0, 5.35, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.27,\"end\":2.62},{\"word\":\"two\",\"start\":5.0,\"end\":5.35}]"} {"utterance_id": "4077-13754-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0007.flac", "task_type": "lexical_detection", "query": "recognized", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"recognized\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.12, 5.79, "recognized"]], "response_mode": "offline", "answer": "[{\"word\":\"recognized\",\"start\":5.12,\"end\":5.79}]"} {"utterance_id": "672-122797-0061", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0061.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.06, 1.49, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.06,\"end\":1.49}]"} {"utterance_id": "4077-13751-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0009.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.69, 1.06, "first"], [10.12, 10.44, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":0.69,\"end\":1.06},{\"word\":\"first\",\"start\":10.12,\"end\":10.44}]"} {"utterance_id": "260-123286-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0017.flac", "task_type": "lexical_detection", "query": "shudder", "prompt": "Once the audio has finished, identify every instance of the word \"shudder\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.62, 1.13, "shudder"]], "response_mode": "offline", "answer": "[{\"word\":\"shudder\",\"start\":0.62,\"end\":1.13}]"} {"utterance_id": "7127-75946-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0000.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Do not answer until the audio ends. Then list each instance of a food name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.63, 2.14, "banquet"]], "response_mode": "offline", "answer": "[{\"word\":\"banquet\",\"start\":1.63,\"end\":2.14}]"} {"utterance_id": "3729-6852-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0020.flac", "task_type": "lexical_detection", "query": "italian", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"italian\" with fields word, start, end (seconds). If none, reply [].", "labels": [[5.33, 6.0, "italian"]], "response_mode": "offline", "answer": "[{\"word\":\"italian\",\"start\":5.33,\"end\":6.0}]"} {"utterance_id": "5142-36600-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36600/5142-36600-0001.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.2, 5.85, "naturalists"]], "response_mode": "offline", "answer": "[{\"word\":\"naturalists\",\"start\":5.2,\"end\":5.85}]"} {"utterance_id": "2961-961-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.53, 1.32, "socrates"]], "response_mode": "offline", "answer": "[{\"word\":\"socrates\",\"start\":0.53,\"end\":1.32}]"} {"utterance_id": "672-122797-0066", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0066.flac", "task_type": "lexical_detection", "query": "quantity", "prompt": "When the utterance is complete, report all instances of the word \"quantity\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.79, 2.32, "quantity"]], "response_mode": "offline", "answer": "[{\"word\":\"quantity\",\"start\":1.79,\"end\":2.32}]"} {"utterance_id": "7127-75947-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0009.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.62, 0.98, "quarter"]], "response_mode": "offline", "answer": "[{\"word\":\"quarter\",\"start\":0.62,\"end\":0.98}]"} {"utterance_id": "5142-36377-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0002.flac", "task_type": "lexical_detection", "query": "recalled", "prompt": "After the audio ends, extract all occurrences of the word \"recalled\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.84, 3.38, "recalled"]], "response_mode": "offline", "answer": "[{\"word\":\"recalled\",\"start\":2.84,\"end\":3.38}]"} {"utterance_id": "4446-2275-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0006.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "You will hear a complete utterance. When it ends, list every instance of a family relation you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.75, 2.05, "cousin"]], "response_mode": "offline", "answer": "[{\"word\":\"cousin\",\"start\":1.75,\"end\":2.05}]"} {"utterance_id": "1188-133604-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0044.flac", "task_type": "lexical_counting", "query": "instead", "prompt": "After the audio ends, tally every instance of the word \"instead\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[3.29, 3.67, "one"], [6.04, 6.35, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.29,\"end\":3.67},{\"word\":\"two\",\"start\":6.04,\"end\":6.35}]"} {"utterance_id": "5142-36377-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0022.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.06, 3.79, "mister meadowcroft"]], "response_mode": "offline", "answer": "[{\"word\":\"mister meadowcroft\",\"start\":3.06,\"end\":3.79}]"} {"utterance_id": "5639-40744-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0029.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.43, 6.8, "son"]], "response_mode": "offline", "answer": "[{\"word\":\"son\",\"start\":6.43,\"end\":6.8}]"} {"utterance_id": "2830-3979-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0012.flac", "task_type": "lexical_detection", "query": "forever", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"forever\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.69, 3.23, "forever"]], "response_mode": "offline", "answer": "[{\"word\":\"forever\",\"start\":2.69,\"end\":3.23}]"} {"utterance_id": "2300-131720-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0022.flac", "task_type": "semantic_detection", "query": "country", "prompt": "You will hear a complete utterance. When it ends, list every instance of a country name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.14, 5.53, "england"]], "response_mode": "offline", "answer": "[{\"word\":\"england\",\"start\":5.14,\"end\":5.53}]"} {"utterance_id": "1188-133604-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0000.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.36, 2.64, "four"]], "response_mode": "offline", "answer": "[{\"word\":\"four\",\"start\":2.36,\"end\":2.64}]"} {"utterance_id": "1284-134647-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0000.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.78, 2.38, "clergy"]], "response_mode": "offline", "answer": "[{\"word\":\"clergy\",\"start\":1.78,\"end\":2.38}]"} {"utterance_id": "8463-294825-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0012.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "Do not answer until the audio ends. Then list each instance of a disease name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.88, 5.55, "depression"]], "response_mode": "offline", "answer": "[{\"word\":\"depression\",\"start\":4.88,\"end\":5.55}]"} {"utterance_id": "2094-142345-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0033.flac", "task_type": "lexical_counting", "query": "i'm", "prompt": "Process the full audio offline. For every occurrence of the word \"i'm\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.98, 8.15, "one"], [10.91, 11.03, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.98,\"end\":8.15},{\"word\":\"two\",\"start\":10.91,\"end\":11.03}]"} {"utterance_id": "5683-32866-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0016.flac", "task_type": "lexical_detection", "query": "aye", "prompt": "After listening to the whole recording, output every instance of the word \"aye\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.39, 3.88, "aye"]], "response_mode": "offline", "answer": "[{\"word\":\"aye\",\"start\":3.39,\"end\":3.88}]"} {"utterance_id": "4992-41806-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0017.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "Once the audio has finished, list each occurrence of a plant name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[1.72, 2.21, "one"], [3.98, 4.76, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.72,\"end\":2.21},{\"word\":\"two\",\"start\":3.98,\"end\":4.76}]"} {"utterance_id": "2094-142345-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0023.flac", "task_type": "lexical_detection", "query": "you're", "prompt": "After the audio ends, extract all occurrences of the word \"you're\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.22, 0.59, "you're"]], "response_mode": "offline", "answer": "[{\"word\":\"you're\",\"start\":0.22,\"end\":0.59}]"} {"utterance_id": "6930-81414-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0026.flac", "task_type": "lexical_detection", "query": "nerves", "prompt": "When the utterance is complete, report all instances of the word \"nerves\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.36, 1.84, "nerves"]], "response_mode": "offline", "answer": "[{\"word\":\"nerves\",\"start\":1.36,\"end\":1.84}]"} {"utterance_id": "2830-3980-0067", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0067.flac", "task_type": "lexical_detection", "query": "amends", "prompt": "After the audio ends, extract all occurrences of the word \"amends\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.39, 6.81, "amends"]], "response_mode": "offline", "answer": "[{\"word\":\"amends\",\"start\":6.39,\"end\":6.81}]"} {"utterance_id": "8463-287645-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0009.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.57, 1.95, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.57,\"end\":1.95}]"} {"utterance_id": "6930-75918-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0011.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.52, 2.94, "count"]], "response_mode": "offline", "answer": "[{\"word\":\"count\",\"start\":2.52,\"end\":2.94}]"} {"utterance_id": "5639-40744-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0019.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Once the audio has finished, identify every instance of a family relation. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.04, 2.47, "child"]], "response_mode": "offline", "answer": "[{\"word\":\"child\",\"start\":2.04,\"end\":2.47}]"} {"utterance_id": "1188-133604-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0021.flac", "task_type": "lexical_counting", "query": "felicity", "prompt": "After the audio ends, tally every instance of the word \"felicity\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[4.0, 4.46, "one"], [10.96, 11.58, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.0,\"end\":4.46},{\"word\":\"two\",\"start\":10.96,\"end\":11.58}]"} {"utterance_id": "3575-170457-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0024.flac", "task_type": "lexical_detection", "query": "amply", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"amply\" with fields word, start, end (seconds). If none, reply [].", "labels": [[9.47, 9.9, "amply"]], "response_mode": "offline", "answer": "[{\"word\":\"amply\",\"start\":9.47,\"end\":9.9}]"} {"utterance_id": "1188-133604-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0023.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "When the utterance is complete, report how an animal name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.63, 5.14, "one"], [7.96, 8.33, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.63,\"end\":5.14},{\"word\":\"two\",\"start\":7.96,\"end\":8.33}]"} {"utterance_id": "260-123286-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0027.flac", "task_type": "lexical_detection", "query": "distinguish", "prompt": "After the audio ends, extract all occurrences of the word \"distinguish\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.5, 0.98, "distinguish"]], "response_mode": "offline", "answer": "[{\"word\":\"distinguish\",\"start\":0.5,\"end\":0.98}]"} {"utterance_id": "1089-134686-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0017.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the entire clip, then report the count sequence for a person's name as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.5, 1.06, "one"], [6.38, 6.8, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.5,\"end\":1.06},{\"word\":\"two\",\"start\":6.38,\"end\":6.8}]"} {"utterance_id": "4992-41797-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0014.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[6.32, 6.9, "bones"]], "response_mode": "offline", "answer": "[{\"word\":\"bones\",\"start\":6.32,\"end\":6.9}]"} {"utterance_id": "3575-170457-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0027.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.19, 3.63, "southey's"]], "response_mode": "offline", "answer": "[{\"word\":\"southey's\",\"start\":3.19,\"end\":3.63}]"} {"utterance_id": "7729-102255-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0014.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.38, 0.91, "leaders"]], "response_mode": "offline", "answer": "[{\"word\":\"leaders\",\"start\":0.38,\"end\":0.91}]"} {"utterance_id": "8463-294828-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0030.flac", "task_type": "lexical_detection", "query": "commander", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"commander\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.8, 1.46, "commander"]], "response_mode": "offline", "answer": "[{\"word\":\"commander\",\"start\":0.8,\"end\":1.46}]"} {"utterance_id": "4077-13751-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0012.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a profession (word, start, end). If none, reply [].", "labels": [[0.49, 1.48, "one"], [3.86, 4.39, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.49,\"end\":1.48},{\"word\":\"two\",\"start\":3.86,\"end\":4.39}]"} {"utterance_id": "4446-2273-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0035.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "When the utterance is complete, report how a body part accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.11, 1.48, "one"], [2.94, 3.28, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.11,\"end\":1.48},{\"word\":\"two\",\"start\":2.94,\"end\":3.28}]"} {"utterance_id": "6829-68769-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0016.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.28, 3.66, "tom"]], "response_mode": "offline", "answer": "[{\"word\":\"tom\",\"start\":3.28,\"end\":3.66}]"} {"utterance_id": "7729-102255-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0017.flac", "task_type": "lexical_detection", "query": "jones", "prompt": "Once the audio has finished, identify every instance of the word \"jones\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.68, 1.12, "jones"]], "response_mode": "offline", "answer": "[{\"word\":\"jones\",\"start\":0.68,\"end\":1.12}]"} {"utterance_id": "1320-122612-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0007.flac", "task_type": "lexical_detection", "query": "bade", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"bade\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.11, 4.4, "bade"]], "response_mode": "offline", "answer": "[{\"word\":\"bade\",\"start\":4.11,\"end\":4.4}]"} {"utterance_id": "4446-2275-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0019.flac", "task_type": "lexical_detection", "query": "can't", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"can't\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.29, 3.69, "can't"]], "response_mode": "offline", "answer": "[{\"word\":\"can't\",\"start\":3.29,\"end\":3.69}]"} {"utterance_id": "8224-274384-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0009.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.05, 4.54, "king"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":4.05,\"end\":4.54}]"} {"utterance_id": "4077-13751-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0004.flac", "task_type": "lexical_detection", "query": "seldom", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"seldom\" with fields word, start, end (seconds). If none, reply [].", "labels": [[10.68, 11.08, "seldom"]], "response_mode": "offline", "answer": "[{\"word\":\"seldom\",\"start\":10.68,\"end\":11.08}]"} {"utterance_id": "61-70968-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0044.flac", "task_type": "lexical_detection", "query": "safe", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"safe\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.98, 1.3, "safe"]], "response_mode": "offline", "answer": "[{\"word\":\"safe\",\"start\":0.98,\"end\":1.3}]"} {"utterance_id": "3575-170457-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0053.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.66, 5.25, "sisters"]], "response_mode": "offline", "answer": "[{\"word\":\"sisters\",\"start\":4.66,\"end\":5.25}]"} {"utterance_id": "3729-6852-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0039.flac", "task_type": "lexical_detection", "query": "politely", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"politely\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.11, 4.67, "politely"]], "response_mode": "offline", "answer": "[{\"word\":\"politely\",\"start\":4.11,\"end\":4.67}]"} {"utterance_id": "7127-75947-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0021.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen carefully to the entire clip, then report every instance of a city name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.89, 3.68, "tonnay charente"]], "response_mode": "offline", "answer": "[{\"word\":\"tonnay charente\",\"start\":2.89,\"end\":3.68}]"} {"utterance_id": "1580-141084-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0046.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.27, 2.13, "mister gilchrist"]], "response_mode": "offline", "answer": "[{\"word\":\"mister gilchrist\",\"start\":1.27,\"end\":2.13}]"} {"utterance_id": "237-134500-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0035.flac", "task_type": "lexical_detection", "query": "cross", "prompt": "After the audio ends, extract all occurrences of the word \"cross\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.2, 1.57, "cross"]], "response_mode": "offline", "answer": "[{\"word\":\"cross\",\"start\":1.2,\"end\":1.57}]"} {"utterance_id": "4446-2271-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0015.flac", "task_type": "lexical_detection", "query": "nerves", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"nerves\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.14, 1.53, "nerves"]], "response_mode": "offline", "answer": "[{\"word\":\"nerves\",\"start\":1.14,\"end\":1.53}]"} {"utterance_id": "2300-131720-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0017.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Process the full audio offline. Find every instance of a plant name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[18.74, 19.07, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":18.74,\"end\":19.07}]"} {"utterance_id": "260-123440-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0008.flac", "task_type": "lexical_counting", "query": "know", "prompt": "Listen to the full audio. Count every occurrence of the word \"know\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.36, 1.8, "one"], [3.14, 3.48, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.36,\"end\":1.8},{\"word\":\"two\",\"start\":3.14,\"end\":3.48}]"} {"utterance_id": "4992-41806-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0000.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Do not answer until the audio ends. Then list each instance of a food name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.45, 5.76, "water"]], "response_mode": "offline", "answer": "[{\"word\":\"water\",\"start\":5.45,\"end\":5.76}]"} {"utterance_id": "61-70968-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0005.flac", "task_type": "lexical_detection", "query": "died", "prompt": "When the utterance is complete, report all instances of the word \"died\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.27, 4.77, "died"]], "response_mode": "offline", "answer": "[{\"word\":\"died\",\"start\":4.27,\"end\":4.77}]"} {"utterance_id": "7729-102255-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0024.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Process the full audio offline. Find every instance of a city name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.19, 2.7, "towns"]], "response_mode": "offline", "answer": "[{\"word\":\"towns\",\"start\":2.19,\"end\":2.7}]"} {"utterance_id": "908-31957-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0015.flac", "task_type": "lexical_detection", "query": "crown", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"crown\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.22, 3.84, "crown"]], "response_mode": "offline", "answer": "[{\"word\":\"crown\",\"start\":3.22,\"end\":3.84}]"} {"utterance_id": "1284-1180-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0003.flac", "task_type": "semantic_detection", "query": "country", "prompt": "You will hear a complete utterance. When it ends, list every instance of a country name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.32, 3.69, "oz"]], "response_mode": "offline", "answer": "[{\"word\":\"oz\",\"start\":3.32,\"end\":3.69}]"} {"utterance_id": "5105-28241-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0007.flac", "task_type": "lexical_detection", "query": "fear", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"fear\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.78, 0.97, "fear"]], "response_mode": "offline", "answer": "[{\"word\":\"fear\",\"start\":0.78,\"end\":0.97}]"} {"utterance_id": "8463-294828-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0031.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "After listening to the whole recording, output timed count words for a profession (one, two, three, ...) as JSON. If none, reply [].", "labels": [[0.7, 1.21, "one"], [5.19, 5.77, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.7,\"end\":1.21},{\"word\":\"two\",\"start\":5.19,\"end\":5.77}]"} {"utterance_id": "1221-135767-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0001.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.77, 4.12, "pair"]], "response_mode": "offline", "answer": "[{\"word\":\"pair\",\"start\":3.77,\"end\":4.12}]"} {"utterance_id": "4970-29093-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0005.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "Listen to the audio in full, then emit a JSON array of count events for an animal name (word, start, end). If none, reply [].", "labels": [[16.76, 17.42, "one"], [19.45, 19.93, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":16.76,\"end\":17.42},{\"word\":\"two\",\"start\":19.45,\"end\":19.93}]"} {"utterance_id": "260-123288-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0026.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.19, 7.8, "hans"]], "response_mode": "offline", "answer": "[{\"word\":\"hans\",\"start\":7.19,\"end\":7.8}]"} {"utterance_id": "5683-32879-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0024.flac", "task_type": "lexical_detection", "query": "love", "prompt": "Process the full audio offline. Find every instance of the word \"love\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.59, 1.84, "love"]], "response_mode": "offline", "answer": "[{\"word\":\"love\",\"start\":1.59,\"end\":1.84}]"} {"utterance_id": "4992-41806-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0009.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.81, 2.16, "wife"]], "response_mode": "offline", "answer": "[{\"word\":\"wife\",\"start\":1.81,\"end\":2.16}]"} {"utterance_id": "4446-2275-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0035.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.48, 1.0, "alexander"]], "response_mode": "offline", "answer": "[{\"word\":\"alexander\",\"start\":0.48,\"end\":1.0}]"} {"utterance_id": "61-70968-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0017.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.52, 4.97, "sons"]], "response_mode": "offline", "answer": "[{\"word\":\"sons\",\"start\":4.52,\"end\":4.97}]"} {"utterance_id": "4992-41797-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0017.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.2, 5.41, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":5.2,\"end\":5.41}]"} {"utterance_id": "7127-75947-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0014.flac", "task_type": "lexical_detection", "query": "completely", "prompt": "After listening to the whole recording, output every instance of the word \"completely\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.08, 8.71, "completely"]], "response_mode": "offline", "answer": "[{\"word\":\"completely\",\"start\":8.08,\"end\":8.71}]"} {"utterance_id": "2961-961-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0022.flac", "task_type": "lexical_counting", "query": "whom", "prompt": "After listening to the whole recording, output timed count words for the word \"whom\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[15.14, 15.37, "one"], [17.43, 17.65, "two"], [23.9, 24.07, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":15.14,\"end\":15.37},{\"word\":\"two\",\"start\":17.43,\"end\":17.65},{\"word\":\"three\",\"start\":23.9,\"end\":24.07}]"} {"utterance_id": "237-126133-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.71, 1.22, "polly"], [2.63, 3.16, "jasper"], [7.08, 7.6, "phronsie"]], "response_mode": "offline", "answer": "[{\"word\":\"polly\",\"start\":0.71,\"end\":1.22},{\"word\":\"jasper\",\"start\":2.63,\"end\":3.16},{\"word\":\"phronsie\",\"start\":7.08,\"end\":7.6}]"} {"utterance_id": "6930-81414-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0019.flac", "task_type": "lexical_detection", "query": "ground", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"ground\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.95, 2.41, "ground"]], "response_mode": "offline", "answer": "[{\"word\":\"ground\",\"start\":1.95,\"end\":2.41}]"} {"utterance_id": "61-70968-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0013.flac", "task_type": "lexical_detection", "query": "terrified", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"terrified\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.62, 4.28, "terrified"]], "response_mode": "offline", "answer": "[{\"word\":\"terrified\",\"start\":3.62,\"end\":4.28}]"} {"utterance_id": "2094-142345-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0024.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.5, 1.09, "munny"]], "response_mode": "offline", "answer": "[{\"word\":\"munny\",\"start\":0.5,\"end\":1.09}]"} {"utterance_id": "6829-68771-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0029.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.18, 0.5, "beth"], [4.47, 4.92, "eliza"]], "response_mode": "offline", "answer": "[{\"word\":\"beth\",\"start\":0.18,\"end\":0.5},{\"word\":\"eliza\",\"start\":4.47,\"end\":4.92}]"} {"utterance_id": "121-123852-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123852/121-123852-0003.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Once the audio has finished, identify every instance of a food name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[8.86, 9.36, "water"]], "response_mode": "offline", "answer": "[{\"word\":\"water\",\"start\":8.86,\"end\":9.36}]"} {"utterance_id": "1580-141084-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0022.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.67, 0.85, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.67,\"end\":0.85}]"} {"utterance_id": "1995-1836-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0013.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.23, 4.59, "ear"]], "response_mode": "offline", "answer": "[{\"word\":\"ear\",\"start\":4.23,\"end\":4.59}]"} {"utterance_id": "5683-32865-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0002.flac", "task_type": "lexical_detection", "query": "shoulder", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"shoulder\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.85, 2.45, "shoulder"]], "response_mode": "offline", "answer": "[{\"word\":\"shoulder\",\"start\":1.85,\"end\":2.45}]"} {"utterance_id": "4992-41806-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0005.flac", "task_type": "lexical_detection", "query": "olive's", "prompt": "After the audio ends, extract all occurrences of the word \"olive's\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.26, 1.76, "olive's"]], "response_mode": "offline", "answer": "[{\"word\":\"olive's\",\"start\":1.26,\"end\":1.76}]"} {"utterance_id": "672-122797-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0038.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.54, 1.25, "fir tree"]], "response_mode": "offline", "answer": "[{\"word\":\"fir tree\",\"start\":0.54,\"end\":1.25}]"} {"utterance_id": "61-70970-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0003.flac", "task_type": "lexical_detection", "query": "beggar", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"beggar\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.25, 1.58, "beggar"]], "response_mode": "offline", "answer": "[{\"word\":\"beggar\",\"start\":1.25,\"end\":1.58}]"} {"utterance_id": "4077-13751-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0009.flac", "task_type": "semantic_counting", "query": "ordinal", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for an ordinal number with word/start/end. If none, reply [].", "labels": [[0.69, 1.06, "one"], [10.12, 10.44, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.69,\"end\":1.06},{\"word\":\"two\",\"start\":10.12,\"end\":10.44}]"} {"utterance_id": "5105-28241-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0011.flac", "task_type": "lexical_detection", "query": "sped", "prompt": "After the audio ends, extract all occurrences of the word \"sped\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.66, 5.06, "sped"]], "response_mode": "offline", "answer": "[{\"word\":\"sped\",\"start\":4.66,\"end\":5.06}]"} {"utterance_id": "1188-133604-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0007.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[3.74, 3.98, "one"], [6.15, 6.32, "two"], [9.83, 10.01, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.74,\"end\":3.98},{\"word\":\"two\",\"start\":6.15,\"end\":6.32},{\"word\":\"one\",\"start\":9.83,\"end\":10.01}]"} {"utterance_id": "1284-1181-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0008.flac", "task_type": "lexical_detection", "query": "qualities", "prompt": "When the utterance is complete, report all instances of the word \"qualities\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.53, 4.15, "qualities"]], "response_mode": "offline", "answer": "[{\"word\":\"qualities\",\"start\":3.53,\"end\":4.15}]"} {"utterance_id": "4507-16021-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0006.flac", "task_type": "lexical_detection", "query": "slang", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"slang\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.25, 0.87, "slang"]], "response_mode": "offline", "answer": "[{\"word\":\"slang\",\"start\":0.25,\"end\":0.87}]"} {"utterance_id": "672-122797-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0025.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.25, 7.54, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.25,\"end\":7.54}]"} {"utterance_id": "5683-32866-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0023.flac", "task_type": "lexical_detection", "query": "furniture", "prompt": "Once the audio has finished, identify every instance of the word \"furniture\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.57, 1.02, "furniture"]], "response_mode": "offline", "answer": "[{\"word\":\"furniture\",\"start\":0.57,\"end\":1.02}]"} {"utterance_id": "4992-41806-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0015.flac", "task_type": "lexical_detection", "query": "cover", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"cover\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.79, 7.15, "cover"]], "response_mode": "offline", "answer": "[{\"word\":\"cover\",\"start\":6.79,\"end\":7.15}]"} {"utterance_id": "260-123440-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0015.flac", "task_type": "lexical_detection", "query": "hadn't", "prompt": "Once the audio has finished, identify every instance of the word \"hadn't\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.93, 1.24, "hadn't"]], "response_mode": "offline", "answer": "[{\"word\":\"hadn't\",\"start\":0.93,\"end\":1.24}]"} {"utterance_id": "4077-13751-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0003.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.03, 1.34, "single"], [7.27, 7.7, "seventy"], [9.29, 10.02, "seven hundred"]], "response_mode": "offline", "answer": "[{\"word\":\"single\",\"start\":1.03,\"end\":1.34},{\"word\":\"seventy\",\"start\":7.27,\"end\":7.7},{\"word\":\"seven hundred\",\"start\":9.29,\"end\":10.02}]"} {"utterance_id": "5105-28241-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0005.flac", "task_type": "lexical_detection", "query": "algeria", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"algeria\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.1, 5.81, "algeria"]], "response_mode": "offline", "answer": "[{\"word\":\"algeria\",\"start\":5.1,\"end\":5.81}]"} {"utterance_id": "5105-28241-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0002.flac", "task_type": "lexical_detection", "query": "rendered", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"rendered\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.53, 2.85, "rendered"]], "response_mode": "offline", "answer": "[{\"word\":\"rendered\",\"start\":2.53,\"end\":2.85}]"} {"utterance_id": "2094-142345-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0020.flac", "task_type": "lexical_detection", "query": "you'd", "prompt": "After the audio ends, extract all occurrences of the word \"you'd\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.67, 0.79, "you'd"]], "response_mode": "offline", "answer": "[{\"word\":\"you'd\",\"start\":0.67,\"end\":0.79}]"} {"utterance_id": "5639-40744-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0013.flac", "task_type": "lexical_detection", "query": "prince", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"prince\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.25, 4.84, "prince"]], "response_mode": "offline", "answer": "[{\"word\":\"prince\",\"start\":4.25,\"end\":4.84}]"} {"utterance_id": "1188-133604-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0024.flac", "task_type": "lexical_detection", "query": "conviction", "prompt": "Once the audio has finished, identify every instance of the word \"conviction\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.79, 5.51, "conviction"]], "response_mode": "offline", "answer": "[{\"word\":\"conviction\",\"start\":4.79,\"end\":5.51}]"} {"utterance_id": "8463-294828-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0031.flac", "task_type": "lexical_detection", "query": "presence", "prompt": "After listening to the whole recording, output every instance of the word \"presence\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.71, 4.25, "presence"]], "response_mode": "offline", "answer": "[{\"word\":\"presence\",\"start\":3.71,\"end\":4.25}]"} {"utterance_id": "5142-33396-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0037.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.06, 3.5, "farmer"]], "response_mode": "offline", "answer": "[{\"word\":\"farmer\",\"start\":3.06,\"end\":3.5}]"} {"utterance_id": "4077-13751-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0007.flac", "task_type": "lexical_counting", "query": "people", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"people\" (word, start, end). If none, reply [].", "labels": [[1.65, 2.03, "one"], [9.14, 9.6, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.65,\"end\":2.03},{\"word\":\"two\",\"start\":9.14,\"end\":9.6}]"} {"utterance_id": "5683-32879-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0004.flac", "task_type": "lexical_detection", "query": "enables", "prompt": "Process the full audio offline. Find every instance of the word \"enables\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.99, 5.51, "enables"]], "response_mode": "offline", "answer": "[{\"word\":\"enables\",\"start\":4.99,\"end\":5.51}]"} {"utterance_id": "1320-122612-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.7, 4.4, "chingachgook"]], "response_mode": "offline", "answer": "[{\"word\":\"chingachgook\",\"start\":3.7,\"end\":4.4}]"} {"utterance_id": "3729-6852-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0033.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the full audio. After it finishes, find every instance of an ordinal number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.81, 10.27, "first"], [15.45, 16.06, "second"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":9.81,\"end\":10.27},{\"word\":\"second\",\"start\":15.45,\"end\":16.06}]"} {"utterance_id": "1995-1837-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0012.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.4, 6.72, "foot"]], "response_mode": "offline", "answer": "[{\"word\":\"foot\",\"start\":6.4,\"end\":6.72}]"} {"utterance_id": "6829-68769-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0019.flac", "task_type": "lexical_detection", "query": "reception", "prompt": "After listening to the whole recording, output every instance of the word \"reception\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.31, 1.87, "reception"]], "response_mode": "offline", "answer": "[{\"word\":\"reception\",\"start\":1.31,\"end\":1.87}]"} {"utterance_id": "8224-274381-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0003.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.01, 1.48, "ears"]], "response_mode": "offline", "answer": "[{\"word\":\"ears\",\"start\":1.01,\"end\":1.48}]"} {"utterance_id": "3570-5695-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0005.flac", "task_type": "lexical_counting", "query": "virtually", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"virtually\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[11.47, 11.98, "one"], [16.07, 16.56, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":11.47,\"end\":11.98},{\"word\":\"two\",\"start\":16.07,\"end\":16.56}]"} {"utterance_id": "7021-79740-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0000.flac", "task_type": "lexical_detection", "query": "training", "prompt": "After listening to the whole recording, output every instance of the word \"training\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.86, 3.28, "training"]], "response_mode": "offline", "answer": "[{\"word\":\"training\",\"start\":2.86,\"end\":3.28}]"} {"utterance_id": "260-123440-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0010.flac", "task_type": "lexical_counting", "query": "how", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"how\" with word/start/end. If none, reply [].", "labels": [[0.14, 0.41, "one"], [2.45, 2.66, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.14,\"end\":0.41},{\"word\":\"two\",\"start\":2.45,\"end\":2.66}]"} {"utterance_id": "1995-1826-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0020.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the full audio. After it finishes, find every instance of a plant name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.85, "cotton"]], "response_mode": "offline", "answer": "[{\"word\":\"cotton\",\"start\":2.35,\"end\":2.85}]"} {"utterance_id": "1284-1180-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0016.flac", "task_type": "lexical_detection", "query": "seemed", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"seemed\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.98, 1.2, "seemed"]], "response_mode": "offline", "answer": "[{\"word\":\"seemed\",\"start\":0.98,\"end\":1.2}]"} {"utterance_id": "1221-135766-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0003.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 0.84, "child"]], "response_mode": "offline", "answer": "[{\"word\":\"child\",\"start\":0.42,\"end\":0.84}]"} {"utterance_id": "5142-33396-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0020.flac", "task_type": "lexical_counting", "query": "smells", "prompt": "Listen to the full audio. Count every occurrence of the word \"smells\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.59, 0.89, "one"], [2.04, 2.38, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.59,\"end\":0.89},{\"word\":\"two\",\"start\":2.04,\"end\":2.38}]"} {"utterance_id": "908-157963-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0006.flac", "task_type": "lexical_detection", "query": "death", "prompt": "After the audio ends, extract all occurrences of the word \"death\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.32, 2.78, "death"]], "response_mode": "offline", "answer": "[{\"word\":\"death\",\"start\":2.32,\"end\":2.78}]"} {"utterance_id": "2300-131720-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0041.flac", "task_type": "lexical_detection", "query": "liquid", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"liquid\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.96, 3.31, "liquid"]], "response_mode": "offline", "answer": "[{\"word\":\"liquid\",\"start\":2.96,\"end\":3.31}]"} {"utterance_id": "4970-29093-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0010.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[10.3, 10.73, "editor"]], "response_mode": "offline", "answer": "[{\"word\":\"editor\",\"start\":10.3,\"end\":10.73}]"} {"utterance_id": "2300-131720-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0003.flac", "task_type": "lexical_detection", "query": "dynamo", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"dynamo\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.67, 1.26, "dynamo"]], "response_mode": "offline", "answer": "[{\"word\":\"dynamo\",\"start\":0.67,\"end\":1.26}]"} {"utterance_id": "1284-1180-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0003.flac", "task_type": "lexical_detection", "query": "explore", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"explore\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.79, 2.33, "explore"]], "response_mode": "offline", "answer": "[{\"word\":\"explore\",\"start\":1.79,\"end\":2.33}]"} {"utterance_id": "5639-40744-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0002.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.44, 2.91, "faces"]], "response_mode": "offline", "answer": "[{\"word\":\"faces\",\"start\":2.44,\"end\":2.91}]"} {"utterance_id": "7176-92135-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0012.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.39, 4.8, "player"], [15.17, 15.78, "dramatist's"]], "response_mode": "offline", "answer": "[{\"word\":\"player\",\"start\":4.39,\"end\":4.8},{\"word\":\"dramatist's\",\"start\":15.17,\"end\":15.78}]"} {"utterance_id": "7176-92135-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0020.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a number with word/start/end. If none, reply [].", "labels": [[0.72, 1.98, "one"], [3.6, 3.81, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.72,\"end\":1.98},{\"word\":\"two\",\"start\":3.6,\"end\":3.81}]"} {"utterance_id": "5142-33396-0049", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0049.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.76, 3.2, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":2.76,\"end\":3.2}]"} {"utterance_id": "1995-1826-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0023.flac", "task_type": "semantic_detection", "query": "food", "prompt": "After listening to the whole recording, output every instance of a food name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.29, 0.87, "goobers"], [5.59, 6.37, "yams"]], "response_mode": "offline", "answer": "[{\"word\":\"goobers\",\"start\":0.29,\"end\":0.87},{\"word\":\"yams\",\"start\":5.59,\"end\":6.37}]"} {"utterance_id": "6930-75918-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0004.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "You will hear a complete utterance. When it ends, list every instance of a family relation you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.07, 1.6, "daughter"]], "response_mode": "offline", "answer": "[{\"word\":\"daughter\",\"start\":1.07,\"end\":1.6}]"} {"utterance_id": "1221-135766-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0009.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.78, 4.24, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":3.78,\"end\":4.24}]"} {"utterance_id": "7176-88083-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0019.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "After the audio ends, tally every instance of a body part. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[2.7, 3.25, "one"], [5.21, 5.62, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.7,\"end\":3.25},{\"word\":\"two\",\"start\":5.21,\"end\":5.62}]"} {"utterance_id": "8455-210777-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0050.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.49, 0.67, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.49,\"end\":0.67}]"} {"utterance_id": "1995-1826-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0024.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After the audio ends, extract all occurrences of a color name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.34, 0.87, "golden"], [2.42, 2.85, "silver"]], "response_mode": "offline", "answer": "[{\"word\":\"golden\",\"start\":0.34,\"end\":0.87},{\"word\":\"silver\",\"start\":2.42,\"end\":2.85}]"} {"utterance_id": "6829-68771-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0003.flac", "task_type": "lexical_detection", "query": "committee", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"committee\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.28, 1.77, "committee"]], "response_mode": "offline", "answer": "[{\"word\":\"committee\",\"start\":1.28,\"end\":1.77}]"} {"utterance_id": "1320-122612-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0015.flac", "task_type": "lexical_detection", "query": "crowded", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"crowded\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.92, 1.35, "crowded"]], "response_mode": "offline", "answer": "[{\"word\":\"crowded\",\"start\":0.92,\"end\":1.35}]"} {"utterance_id": "8463-294828-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0027.flac", "task_type": "lexical_detection", "query": "plants", "prompt": "Process the full audio offline. Find every instance of the word \"plants\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.18, 4.63, "plants"]], "response_mode": "offline", "answer": "[{\"word\":\"plants\",\"start\":4.18,\"end\":4.63}]"} {"utterance_id": "61-70968-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.41, 1.6, "sister nell"]], "response_mode": "offline", "answer": "[{\"word\":\"sister nell\",\"start\":0.41,\"end\":1.6}]"} {"utterance_id": "1580-141083-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0019.flac", "task_type": "lexical_detection", "query": "students", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"students\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.97, 1.46, "students"]], "response_mode": "offline", "answer": "[{\"word\":\"students\",\"start\":0.97,\"end\":1.46}]"} {"utterance_id": "8463-294828-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0033.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.94, 5.53, "officers"]], "response_mode": "offline", "answer": "[{\"word\":\"officers\",\"start\":4.94,\"end\":5.53}]"} {"utterance_id": "672-122797-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0021.flac", "task_type": "lexical_detection", "query": "christmas", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"christmas\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.85, 1.37, "christmas"]], "response_mode": "offline", "answer": "[{\"word\":\"christmas\",\"start\":0.85,\"end\":1.37}]"} {"utterance_id": "4970-29095-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0011.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.31, 2.59, "six"]], "response_mode": "offline", "answer": "[{\"word\":\"six\",\"start\":2.31,\"end\":2.59}]"} {"utterance_id": "4446-2273-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0005.flac", "task_type": "lexical_detection", "query": "jolly", "prompt": "Process the full audio offline. Find every instance of the word \"jolly\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.27, 2.66, "jolly"]], "response_mode": "offline", "answer": "[{\"word\":\"jolly\",\"start\":2.27,\"end\":2.66}]"} {"utterance_id": "8463-294828-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0014.flac", "task_type": "lexical_detection", "query": "world's", "prompt": "Process the full audio offline. Find every instance of the word \"world's\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.24, 3.67, "world's"]], "response_mode": "offline", "answer": "[{\"word\":\"world's\",\"start\":3.24,\"end\":3.67}]"} {"utterance_id": "2961-960-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0014.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "After listening to the whole recording, output every instance of an animal name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.69, 8.61, "animals birds fishes"]], "response_mode": "offline", "answer": "[{\"word\":\"animals birds fishes\",\"start\":6.69,\"end\":8.61}]"} {"utterance_id": "1284-1180-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0008.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "When the utterance is complete, report all instances of an ordinal number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.91, 7.3, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":6.91,\"end\":7.3}]"} {"utterance_id": "4507-16021-0049", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0049.flac", "task_type": "lexical_detection", "query": "wholly", "prompt": "When the utterance is complete, report all instances of the word \"wholly\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.66, 4.09, "wholly"]], "response_mode": "offline", "answer": "[{\"word\":\"wholly\",\"start\":3.66,\"end\":4.09}]"} {"utterance_id": "4507-16021-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0042.flac", "task_type": "semantic_counting", "query": "color", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a color name: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.77, 1.41, "one"], [3.34, 3.84, "two"], [6.72, 7.85, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.77,\"end\":1.41},{\"word\":\"two\",\"start\":3.34,\"end\":3.84},{\"word\":\"three\",\"start\":6.72,\"end\":7.85}]"} {"utterance_id": "6930-75918-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0018.flac", "task_type": "lexical_detection", "query": "tempest", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"tempest\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.35, 3.81, "tempest"]], "response_mode": "offline", "answer": "[{\"word\":\"tempest\",\"start\":3.35,\"end\":3.81}]"} {"utterance_id": "2300-131720-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0039.flac", "task_type": "lexical_detection", "query": "problem", "prompt": "After the audio ends, extract all occurrences of the word \"problem\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.44, 1.12, "problem"]], "response_mode": "offline", "answer": "[{\"word\":\"problem\",\"start\":0.44,\"end\":1.12}]"} {"utterance_id": "5683-32879-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.46, 0.92, "rachel's"]], "response_mode": "offline", "answer": "[{\"word\":\"rachel's\",\"start\":0.46,\"end\":0.92}]"} {"utterance_id": "2830-3980-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0018.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, list every instance of an organization name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.47, 3.01, "ministry"]], "response_mode": "offline", "answer": "[{\"word\":\"ministry\",\"start\":2.47,\"end\":3.01}]"} {"utterance_id": "7127-75947-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0037.flac", "task_type": "lexical_detection", "query": "seriously", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"seriously\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.28, 2.02, "seriously"]], "response_mode": "offline", "answer": "[{\"word\":\"seriously\",\"start\":1.28,\"end\":2.02}]"} {"utterance_id": "4970-29093-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0007.flac", "task_type": "lexical_detection", "query": "ambition", "prompt": "Once the audio has finished, identify every instance of the word \"ambition\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.59, 2.22, "ambition"]], "response_mode": "offline", "answer": "[{\"word\":\"ambition\",\"start\":1.59,\"end\":2.22}]"} {"utterance_id": "121-127105-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0017.flac", "task_type": "lexical_detection", "query": "tone", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"tone\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.8, 1.05, "tone"]], "response_mode": "offline", "answer": "[{\"word\":\"tone\",\"start\":0.8,\"end\":1.05}]"} {"utterance_id": "4970-29095-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0012.flac", "task_type": "lexical_detection", "query": "besides", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"besides\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.57, 1.42, "besides"]], "response_mode": "offline", "answer": "[{\"word\":\"besides\",\"start\":0.57,\"end\":1.42}]"} {"utterance_id": "6829-68771-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[11.75, 12.58, "mister hopkins"]], "response_mode": "offline", "answer": "[{\"word\":\"mister hopkins\",\"start\":11.75,\"end\":12.58}]"} {"utterance_id": "2094-142345-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0006.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a disease name with fields word, start, end (seconds). If none, reply [].", "labels": [[5.54, 6.17, "mutilation"]], "response_mode": "offline", "answer": "[{\"word\":\"mutilation\",\"start\":5.54,\"end\":6.17}]"} {"utterance_id": "3570-5694-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0006.flac", "task_type": "lexical_detection", "query": "tend", "prompt": "After the audio ends, extract all occurrences of the word \"tend\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.22, 5.62, "tend"]], "response_mode": "offline", "answer": "[{\"word\":\"tend\",\"start\":5.22,\"end\":5.62}]"} {"utterance_id": "8224-274384-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0002.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "When the utterance is complete, report all instances of an organization name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.65, 2.74, "english parliament"]], "response_mode": "offline", "answer": "[{\"word\":\"english parliament\",\"start\":1.65,\"end\":2.74}]"} {"utterance_id": "7127-75946-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0023.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[0.67, 0.97, "king"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":0.67,\"end\":0.97}]"} {"utterance_id": "1284-1181-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0014.flac", "task_type": "semantic_detection", "query": "food", "prompt": "After listening to the whole recording, output every instance of a food name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.56, 2.88, "pepper"]], "response_mode": "offline", "answer": "[{\"word\":\"pepper\",\"start\":2.56,\"end\":2.88}]"} {"utterance_id": "4992-41806-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0017.flac", "task_type": "lexical_detection", "query": "corn", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"corn\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.98, 4.33, "corn"]], "response_mode": "offline", "answer": "[{\"word\":\"corn\",\"start\":3.98,\"end\":4.33}]"} {"utterance_id": "3570-5694-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0020.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "Listen to the full audio. Count every occurrence of a profession. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.33, 1.8, "one"], [3.15, 3.7, "two"], [13.15, 13.89, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.33,\"end\":1.8},{\"word\":\"two\",\"start\":3.15,\"end\":3.7},{\"word\":\"three\",\"start\":13.15,\"end\":13.89}]"} {"utterance_id": "2300-131720-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0001.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Listen carefully to the entire clip, then report every instance of a state or province name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.13, 13.66, "holborn"]], "response_mode": "offline", "answer": "[{\"word\":\"holborn\",\"start\":13.13,\"end\":13.66}]"} {"utterance_id": "4970-29093-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0015.flac", "task_type": "lexical_detection", "query": "carrying", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"carrying\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.06, 1.5, "carrying"]], "response_mode": "offline", "answer": "[{\"word\":\"carrying\",\"start\":1.06,\"end\":1.5}]"} {"utterance_id": "7729-102255-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0002.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[7.04, 7.25, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":7.04,\"end\":7.25}]"} {"utterance_id": "2300-131720-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0034.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen to the full audio. After it finishes, find every instance of a money or currency term. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.34, 7.8, "dividend"]], "response_mode": "offline", "answer": "[{\"word\":\"dividend\",\"start\":7.34,\"end\":7.8}]"} {"utterance_id": "3575-170457-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0046.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.36, 2.15, "brontes"], [24.03, 24.47, "tabby's"]], "response_mode": "offline", "answer": "[{\"word\":\"brontes\",\"start\":1.36,\"end\":2.15},{\"word\":\"tabby's\",\"start\":24.03,\"end\":24.47}]"} {"utterance_id": "1995-1837-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0024.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a plant name with fields word, start, end (seconds). If none, reply [].", "labels": [[4.63, 5.16, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":4.63,\"end\":5.16}]"} {"utterance_id": "7729-102255-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0044.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After listening to the whole recording, output every instance of an organization name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.94, 5.09, "kickapoo rangers"]], "response_mode": "offline", "answer": "[{\"word\":\"kickapoo rangers\",\"start\":3.94,\"end\":5.09}]"} {"utterance_id": "7021-79740-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0001.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.52, 0.92, "della"], [2.66, 3.13, "maria"], [5.26, 5.79, "jane"]], "response_mode": "offline", "answer": "[{\"word\":\"della\",\"start\":0.52,\"end\":0.92},{\"word\":\"maria\",\"start\":2.66,\"end\":3.13},{\"word\":\"jane\",\"start\":5.26,\"end\":5.79}]"} {"utterance_id": "7127-75947-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0016.flac", "task_type": "lexical_detection", "query": "mademoiselle", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"mademoiselle\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.75, 1.66, "mademoiselle"]], "response_mode": "offline", "answer": "[{\"word\":\"mademoiselle\",\"start\":0.75,\"end\":1.66}]"} {"utterance_id": "1995-1837-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0027.flac", "task_type": "lexical_detection", "query": "island", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"island\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.7, 4.24, "island"]], "response_mode": "offline", "answer": "[{\"word\":\"island\",\"start\":3.7,\"end\":4.24}]"} {"utterance_id": "7127-75946-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0000.flac", "task_type": "lexical_detection", "query": "imagination", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"imagination\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[13.17, 14.02, "imagination"]], "response_mode": "offline", "answer": "[{\"word\":\"imagination\",\"start\":13.17,\"end\":14.02}]"} {"utterance_id": "2830-3980-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0007.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.74, 1.27, "ambassador"], [5.52, 5.94, "minister"]], "response_mode": "offline", "answer": "[{\"word\":\"ambassador\",\"start\":0.74,\"end\":1.27},{\"word\":\"minister\",\"start\":5.52,\"end\":5.94}]"} {"utterance_id": "4446-2275-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0023.flac", "task_type": "lexical_detection", "query": "somehow", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"somehow\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.73, 3.12, "somehow"]], "response_mode": "offline", "answer": "[{\"word\":\"somehow\",\"start\":2.73,\"end\":3.12}]"} {"utterance_id": "4507-16021-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0029.flac", "task_type": "lexical_detection", "query": "word", "prompt": "When the utterance is complete, report all instances of the word \"word\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.46, 2.86, "word"]], "response_mode": "offline", "answer": "[{\"word\":\"word\",\"start\":2.46,\"end\":2.86}]"} {"utterance_id": "8230-279154-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0019.flac", "task_type": "lexical_detection", "query": "recent", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"recent\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.22, 2.77, "recent"]], "response_mode": "offline", "answer": "[{\"word\":\"recent\",\"start\":2.22,\"end\":2.77}]"} {"utterance_id": "1995-1826-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0001.flac", "task_type": "lexical_detection", "query": "climate", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"climate\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.41, 7.0, "climate"]], "response_mode": "offline", "answer": "[{\"word\":\"climate\",\"start\":6.41,\"end\":7.0}]"} {"utterance_id": "7176-92135-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0036.flac", "task_type": "lexical_detection", "query": "hero", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"hero\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.08, 2.45, "hero"]], "response_mode": "offline", "answer": "[{\"word\":\"hero\",\"start\":2.08,\"end\":2.45}]"} {"utterance_id": "908-157963-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0019.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.25, 12.09, "flowers"]], "response_mode": "offline", "answer": "[{\"word\":\"flowers\",\"start\":11.25,\"end\":12.09}]"} {"utterance_id": "4970-29093-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0001.flac", "task_type": "semantic_detection", "query": "money", "prompt": "After listening to the whole recording, output every instance of a money or currency term with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.1, 4.7, "fortune"]], "response_mode": "offline", "answer": "[{\"word\":\"fortune\",\"start\":4.1,\"end\":4.7}]"} {"utterance_id": "8555-292519-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0001.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Process the full audio offline. Find every instance of a plant name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.82, 10.84, "hollyhocks"], [11.93, 12.49, "poppies"], [15.48, 16.01, "birches"], [19.22, 20.33, "lily plants"]], "response_mode": "offline", "answer": "[{\"word\":\"hollyhocks\",\"start\":9.82,\"end\":10.84},{\"word\":\"poppies\",\"start\":11.93,\"end\":12.49},{\"word\":\"birches\",\"start\":15.48,\"end\":16.01},{\"word\":\"lily plants\",\"start\":19.22,\"end\":20.33}]"} {"utterance_id": "6829-68769-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.34, 2.74, "lucy"]], "response_mode": "offline", "answer": "[{\"word\":\"lucy\",\"start\":2.34,\"end\":2.74}]"} {"utterance_id": "4446-2273-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0019.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen to the full audio. After it finishes, find every instance of a food name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.28, 1.62, "coffee"]], "response_mode": "offline", "answer": "[{\"word\":\"coffee\",\"start\":1.28,\"end\":1.62}]"} {"utterance_id": "672-122797-0073", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0073.flac", "task_type": "lexical_detection", "query": "flamed", "prompt": "After the audio ends, extract all occurrences of the word \"flamed\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.97, 1.63, "flamed"]], "response_mode": "offline", "answer": "[{\"word\":\"flamed\",\"start\":0.97,\"end\":1.63}]"} {"utterance_id": "6829-68771-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0018.flac", "task_type": "lexical_detection", "query": "composure", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"composure\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.58, 5.34, "composure"]], "response_mode": "offline", "answer": "[{\"word\":\"composure\",\"start\":4.58,\"end\":5.34}]"} {"utterance_id": "4507-16021-0055", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0055.flac", "task_type": "lexical_detection", "query": "reading", "prompt": "When the utterance is complete, report all instances of the word \"reading\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.67, 1.24, "reading"]], "response_mode": "offline", "answer": "[{\"word\":\"reading\",\"start\":0.67,\"end\":1.24}]"} {"utterance_id": "7176-92135-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0005.flac", "task_type": "lexical_detection", "query": "suppose", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"suppose\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.4, 0.76, "suppose"]], "response_mode": "offline", "answer": "[{\"word\":\"suppose\",\"start\":0.4,\"end\":0.76}]"} {"utterance_id": "1284-1180-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0017.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.32, 0.56, "one"], [6.62, 6.99, "four"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.32,\"end\":0.56},{\"word\":\"four\",\"start\":6.62,\"end\":6.99}]"} {"utterance_id": "5683-32879-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0011.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.92, 6.28, "cheek"], [7.64, 8.05, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"cheek\",\"start\":5.92,\"end\":6.28},{\"word\":\"eyes\",\"start\":7.64,\"end\":8.05}]"} {"utterance_id": "61-70968-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0027.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.21, 0.56, "robin"]], "response_mode": "offline", "answer": "[{\"word\":\"robin\",\"start\":0.21,\"end\":0.56}]"} {"utterance_id": "908-31957-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0012.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.5, 1.85, "one"], [3.09, 3.39, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.5,\"end\":1.85},{\"word\":\"one\",\"start\":3.09,\"end\":3.39}]"} {"utterance_id": "1221-135766-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0002.flac", "task_type": "lexical_detection", "query": "affected", "prompt": "Process the full audio offline. Find every instance of the word \"affected\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.26, 1.66, "affected"]], "response_mode": "offline", "answer": "[{\"word\":\"affected\",\"start\":1.26,\"end\":1.66}]"} {"utterance_id": "2830-3979-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0009.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "After the audio ends, tally every instance of a person's name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[1.75, 2.12, "one"], [6.69, 7.08, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.75,\"end\":2.12},{\"word\":\"two\",\"start\":6.69,\"end\":7.08}]"} {"utterance_id": "5142-33396-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0010.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.32, 1.68, "tail"], [2.81, 3.23, "head"]], "response_mode": "offline", "answer": "[{\"word\":\"tail\",\"start\":1.32,\"end\":1.68},{\"word\":\"head\",\"start\":2.81,\"end\":3.23}]"} {"utterance_id": "4446-2275-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0000.flac", "task_type": "lexical_detection", "query": "impatience", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"impatience\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.47, 6.18, "impatience"]], "response_mode": "offline", "answer": "[{\"word\":\"impatience\",\"start\":5.47,\"end\":6.18}]"} {"utterance_id": "8455-210777-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0028.flac", "task_type": "lexical_detection", "query": "become", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"become\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.03, 1.42, "become"]], "response_mode": "offline", "answer": "[{\"word\":\"become\",\"start\":1.03,\"end\":1.42}]"} {"utterance_id": "3570-5695-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0014.flac", "task_type": "lexical_detection", "query": "comforts", "prompt": "After the audio ends, extract all occurrences of the word \"comforts\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.19, 4.78, "comforts"]], "response_mode": "offline", "answer": "[{\"word\":\"comforts\",\"start\":4.19,\"end\":4.78}]"} {"utterance_id": "7729-102255-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0012.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.23, 1.03, "several hundred"]], "response_mode": "offline", "answer": "[{\"word\":\"several hundred\",\"start\":0.23,\"end\":1.03}]"} {"utterance_id": "1995-1826-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.51, 2.09, "john"]], "response_mode": "offline", "answer": "[{\"word\":\"john\",\"start\":1.51,\"end\":2.09}]"} {"utterance_id": "4992-23283-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0014.flac", "task_type": "lexical_detection", "query": "mean", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"mean\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.27, 1.49, "mean"]], "response_mode": "offline", "answer": "[{\"word\":\"mean\",\"start\":1.27,\"end\":1.49}]"} {"utterance_id": "5142-33396-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0008.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen carefully to the entire clip, then report every instance of a color name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.06, 1.32, "red"]], "response_mode": "offline", "answer": "[{\"word\":\"red\",\"start\":1.06,\"end\":1.32}]"} {"utterance_id": "3575-170457-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0042.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "Process the full audio offline. Find every instance of a disease name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.24, 1.64, "fracture"]], "response_mode": "offline", "answer": "[{\"word\":\"fracture\",\"start\":1.24,\"end\":1.64}]"} {"utterance_id": "1580-141083-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0021.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.82, 3.21, "guide"]], "response_mode": "offline", "answer": "[{\"word\":\"guide\",\"start\":2.82,\"end\":3.21}]"} {"utterance_id": "5105-28241-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.85, 3.4, "procope"]], "response_mode": "offline", "answer": "[{\"word\":\"procope\",\"start\":2.85,\"end\":3.4}]"} {"utterance_id": "4446-2275-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0020.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.02, 2.46, "hilda"]], "response_mode": "offline", "answer": "[{\"word\":\"hilda\",\"start\":2.02,\"end\":2.46}]"} {"utterance_id": "1995-1837-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0029.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.99, 1.44, "trees"]], "response_mode": "offline", "answer": "[{\"word\":\"trees\",\"start\":0.99,\"end\":1.44}]"} {"utterance_id": "6930-81414-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0019.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.33, 0.85, "voltaire"]], "response_mode": "offline", "answer": "[{\"word\":\"voltaire\",\"start\":0.33,\"end\":0.85}]"} {"utterance_id": "7729-102255-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0043.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.39, 0.6, "few"]], "response_mode": "offline", "answer": "[{\"word\":\"few\",\"start\":0.39,\"end\":0.6}]"} {"utterance_id": "7021-85628-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0003.flac", "task_type": "lexical_counting", "query": "knife", "prompt": "After listening to the whole recording, output timed count words for the word \"knife\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[1.8, 2.16, "one"], [14.35, 14.79, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.8,\"end\":2.16},{\"word\":\"two\",\"start\":14.35,\"end\":14.79}]"} {"utterance_id": "2961-960-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.69, 1.63, "heracleitus"], [5.01, 5.99, "anaxagoras"]], "response_mode": "offline", "answer": "[{\"word\":\"heracleitus\",\"start\":0.69,\"end\":1.63},{\"word\":\"anaxagoras\",\"start\":5.01,\"end\":5.99}]"} {"utterance_id": "260-123286-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0023.flac", "task_type": "lexical_detection", "query": "watery", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"watery\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.74, 2.12, "watery"]], "response_mode": "offline", "answer": "[{\"word\":\"watery\",\"start\":1.74,\"end\":2.12}]"} {"utterance_id": "6829-68769-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0031.flac", "task_type": "lexical_detection", "query": "regards", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"regards\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.67, 1.25, "regards"]], "response_mode": "offline", "answer": "[{\"word\":\"regards\",\"start\":0.67,\"end\":1.25}]"} {"utterance_id": "5639-40744-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0029.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.29, 2.64, "lips"], [3.87, 4.23, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"lips\",\"start\":2.29,\"end\":2.64},{\"word\":\"face\",\"start\":3.87,\"end\":4.23}]"} {"utterance_id": "7127-75946-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0008.flac", "task_type": "lexical_detection", "query": "attempt", "prompt": "After listening to the whole recording, output every instance of the word \"attempt\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.63, 4.07, "attempt"]], "response_mode": "offline", "answer": "[{\"word\":\"attempt\",\"start\":3.63,\"end\":4.07}]"} {"utterance_id": "672-122797-0074", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0074.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Process the full audio offline. Find every instance of a plant name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.53, 4.86, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":4.53,\"end\":4.86}]"} {"utterance_id": "7729-102255-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0007.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.77, 11.41, "auditors"]], "response_mode": "offline", "answer": "[{\"word\":\"auditors\",\"start\":10.77,\"end\":11.41}]"} {"utterance_id": "8230-279154-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0013.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[1.76, 2.34, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":1.76,\"end\":2.34}]"} {"utterance_id": "4992-41797-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0018.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "You will hear a complete utterance. When it ends, list every instance of a plant name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.55, 1.94, "cedar"]], "response_mode": "offline", "answer": "[{\"word\":\"cedar\",\"start\":1.55,\"end\":1.94}]"} {"utterance_id": "4446-2273-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0001.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.91, 4.32, "hilda"]], "response_mode": "offline", "answer": "[{\"word\":\"hilda\",\"start\":3.91,\"end\":4.32}]"} {"utterance_id": "2961-960-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0010.flac", "task_type": "lexical_detection", "query": "defined", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"defined\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.97, 2.7, "defined"]], "response_mode": "offline", "answer": "[{\"word\":\"defined\",\"start\":1.97,\"end\":2.7}]"} {"utterance_id": "1284-1180-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0015.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.9, 2.39, "ojo"]], "response_mode": "offline", "answer": "[{\"word\":\"ojo\",\"start\":1.9,\"end\":2.39}]"} {"utterance_id": "6930-81414-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0021.flac", "task_type": "lexical_detection", "query": "flashed", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"flashed\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.54, 2.21, "flashed"]], "response_mode": "offline", "answer": "[{\"word\":\"flashed\",\"start\":1.54,\"end\":2.21}]"} {"utterance_id": "8224-274381-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0012.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.33, 1.22, "montrose"], [16.92, 17.81, "lord gordon"]], "response_mode": "offline", "answer": "[{\"word\":\"montrose\",\"start\":0.33,\"end\":1.22},{\"word\":\"lord gordon\",\"start\":16.92,\"end\":17.81}]"} {"utterance_id": "4446-2273-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0010.flac", "task_type": "lexical_detection", "query": "stuffed", "prompt": "After listening to the whole recording, output every instance of the word \"stuffed\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.78, 4.04, "stuffed"]], "response_mode": "offline", "answer": "[{\"word\":\"stuffed\",\"start\":3.78,\"end\":4.04}]"} {"utterance_id": "8224-274384-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0000.flac", "task_type": "lexical_detection", "query": "hill", "prompt": "After the audio ends, extract all occurrences of the word \"hill\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.64, 7.22, "hill"]], "response_mode": "offline", "answer": "[{\"word\":\"hill\",\"start\":6.64,\"end\":7.22}]"} {"utterance_id": "4992-23283-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0001.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.54, 1.33, "miss milner's"]], "response_mode": "offline", "answer": "[{\"word\":\"miss milner's\",\"start\":0.54,\"end\":1.33}]"} {"utterance_id": "2830-3980-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0027.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen carefully to the entire clip, then report every instance of a family relation as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.74, 1.2, "brethren"]], "response_mode": "offline", "answer": "[{\"word\":\"brethren\",\"start\":0.74,\"end\":1.2}]"} {"utterance_id": "672-122797-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0013.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Process the full audio offline. Find every instance of a plant name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.67, 3.4, "branches"]], "response_mode": "offline", "answer": "[{\"word\":\"branches\",\"start\":2.67,\"end\":3.4}]"} {"utterance_id": "2961-961-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.89, 4.37, "solon"]], "response_mode": "offline", "answer": "[{\"word\":\"solon\",\"start\":3.89,\"end\":4.37}]"} {"utterance_id": "4992-41797-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0021.flac", "task_type": "lexical_detection", "query": "draws", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"draws\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.23, 5.79, "draws"]], "response_mode": "offline", "answer": "[{\"word\":\"draws\",\"start\":5.23,\"end\":5.79}]"} {"utterance_id": "260-123286-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0028.flac", "task_type": "lexical_detection", "query": "jaw", "prompt": "When the utterance is complete, report all instances of the word \"jaw\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.74, 1.09, "jaw"]], "response_mode": "offline", "answer": "[{\"word\":\"jaw\",\"start\":0.74,\"end\":1.09}]"} {"utterance_id": "3570-5694-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0020.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.33, 1.8, "offices"], [3.15, 3.7, "employment"], [13.15, 13.89, "employments"]], "response_mode": "offline", "answer": "[{\"word\":\"offices\",\"start\":1.33,\"end\":1.8},{\"word\":\"employment\",\"start\":3.15,\"end\":3.7},{\"word\":\"employments\",\"start\":13.15,\"end\":13.89}]"} {"utterance_id": "237-126133-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0011.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.91, 5.36, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":4.91,\"end\":5.36}]"} {"utterance_id": "61-70968-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0051.flac", "task_type": "lexical_detection", "query": "sheriff", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"sheriff\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.45, 1.91, "sheriff"]], "response_mode": "offline", "answer": "[{\"word\":\"sheriff\",\"start\":1.45,\"end\":1.91}]"} {"utterance_id": "1089-134686-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0023.flac", "task_type": "lexical_counting", "query": "bread", "prompt": "After the audio ends, tally every instance of the word \"bread\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[4.58, 4.92, "one"], [10.63, 10.92, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.58,\"end\":4.92},{\"word\":\"two\",\"start\":10.63,\"end\":10.92}]"} {"utterance_id": "61-70970-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0008.flac", "task_type": "lexical_detection", "query": "bed", "prompt": "Once the audio has finished, identify every instance of the word \"bed\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.87, 1.27, "bed"]], "response_mode": "offline", "answer": "[{\"word\":\"bed\",\"start\":0.87,\"end\":1.27}]"} {"utterance_id": "4446-2273-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0018.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.2, 1.57, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":1.2,\"end\":1.57}]"} {"utterance_id": "2830-3980-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0023.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.97, 2.42, "christ"]], "response_mode": "offline", "answer": "[{\"word\":\"christ\",\"start\":1.97,\"end\":2.42}]"} {"utterance_id": "260-123288-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0026.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.3, 12.81, "uncle"]], "response_mode": "offline", "answer": "[{\"word\":\"uncle\",\"start\":12.3,\"end\":12.81}]"} {"utterance_id": "61-70968-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0020.flac", "task_type": "semantic_detection", "query": "city", "prompt": "After listening to the whole recording, output every instance of a city name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.25, 4.85, "nottingham"]], "response_mode": "offline", "answer": "[{\"word\":\"nottingham\",\"start\":4.25,\"end\":4.85}]"} {"utterance_id": "7127-75946-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0017.flac", "task_type": "lexical_detection", "query": "legs", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"legs\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 1.16, "legs"]], "response_mode": "offline", "answer": "[{\"word\":\"legs\",\"start\":0.53,\"end\":1.16}]"} {"utterance_id": "7127-75947-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0020.flac", "task_type": "lexical_detection", "query": "dancing", "prompt": "When the utterance is complete, report all instances of the word \"dancing\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.28, 1.92, "dancing"]], "response_mode": "offline", "answer": "[{\"word\":\"dancing\",\"start\":1.28,\"end\":1.92}]"} {"utterance_id": "2830-3980-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0001.flac", "task_type": "lexical_detection", "query": "highly", "prompt": "After listening to the whole recording, output every instance of the word \"highly\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.92, 3.25, "highly"]], "response_mode": "offline", "answer": "[{\"word\":\"highly\",\"start\":2.92,\"end\":3.25}]"} {"utterance_id": "2830-3980-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0023.flac", "task_type": "lexical_detection", "query": "works", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"works\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.66, 5.02, "works"]], "response_mode": "offline", "answer": "[{\"word\":\"works\",\"start\":4.66,\"end\":5.02}]"} {"utterance_id": "4446-2273-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0003.flac", "task_type": "lexical_detection", "query": "square", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"square\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.89, 2.31, "square"]], "response_mode": "offline", "answer": "[{\"word\":\"square\",\"start\":1.89,\"end\":2.31}]"} {"utterance_id": "5105-28241-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0001.flac", "task_type": "lexical_detection", "query": "spend", "prompt": "Once the audio has finished, identify every instance of the word \"spend\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[11.97, 12.35, "spend"]], "response_mode": "offline", "answer": "[{\"word\":\"spend\",\"start\":11.97,\"end\":12.35}]"} {"utterance_id": "4992-41806-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0008.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the entire clip, then report the count sequence for a person's name as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 0.87, "one"], [3.37, 4.09, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.52,\"end\":0.87},{\"word\":\"two\",\"start\":3.37,\"end\":4.09}]"} {"utterance_id": "1221-135767-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0004.flac", "task_type": "lexical_counting", "query": "deep", "prompt": "When the utterance is complete, report how the word \"deep\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.13, 5.47, "one"], [13.75, 14.14, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.13,\"end\":5.47},{\"word\":\"two\",\"start\":13.75,\"end\":14.14}]"} {"utterance_id": "7021-79740-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0008.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[1.8, 2.11, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.8,\"end\":2.11}]"} {"utterance_id": "1221-135767-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0017.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen to the full audio. After it finishes, find every instance of a food name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[15.83, 16.23, "ale"]], "response_mode": "offline", "answer": "[{\"word\":\"ale\",\"start\":15.83,\"end\":16.23}]"} {"utterance_id": "6930-75918-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0003.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "After the audio ends, tally every instance of a profession. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[3.87, 4.35, "one"], [11.8, 12.2, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.87,\"end\":4.35},{\"word\":\"two\",\"start\":11.8,\"end\":12.2}]"} {"utterance_id": "1089-134686-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0003.flac", "task_type": "lexical_detection", "query": "hello", "prompt": "Process the full audio offline. Find every instance of the word \"hello\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.4, 0.69, "hello"]], "response_mode": "offline", "answer": "[{\"word\":\"hello\",\"start\":0.4,\"end\":0.69}]"} {"utterance_id": "61-70968-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0013.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.08, 2.31, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":2.08,\"end\":2.31}]"} {"utterance_id": "5639-40744-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0041.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.87, 1.32, "rodolfo"], [7.14, 7.77, "leocadia"]], "response_mode": "offline", "answer": "[{\"word\":\"rodolfo\",\"start\":0.87,\"end\":1.32},{\"word\":\"leocadia\",\"start\":7.14,\"end\":7.77}]"} {"utterance_id": "4446-2271-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0013.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "When the utterance is complete, report all instances of an ordinal number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.64, 3.91, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":3.64,\"end\":3.91}]"} {"utterance_id": "5639-40744-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0005.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.96, 1.25, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.96,\"end\":1.25}]"} {"utterance_id": "5683-32879-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0015.flac", "task_type": "lexical_detection", "query": "rachel", "prompt": "Process the full audio offline. Find every instance of the word \"rachel\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.21, 1.71, "rachel"]], "response_mode": "offline", "answer": "[{\"word\":\"rachel\",\"start\":1.21,\"end\":1.71}]"} {"utterance_id": "7021-79759-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79759/7021-79759-0000.flac", "task_type": "lexical_detection", "query": "produced", "prompt": "Once the audio has finished, identify every instance of the word \"produced\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.73, 2.4, "produced"]], "response_mode": "offline", "answer": "[{\"word\":\"produced\",\"start\":1.73,\"end\":2.4}]"} {"utterance_id": "1320-122612-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0012.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Process the full audio offline. Find every instance of a food name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.55, 4.02, "offals"]], "response_mode": "offline", "answer": "[{\"word\":\"offals\",\"start\":3.55,\"end\":4.02}]"} {"utterance_id": "1580-141084-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0002.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Once the audio has finished, identify every instance of an organization name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.27, 2.9, "college"]], "response_mode": "offline", "answer": "[{\"word\":\"college\",\"start\":2.27,\"end\":2.9}]"} {"utterance_id": "1284-1181-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0019.flac", "task_type": "lexical_detection", "query": "ornamental", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"ornamental\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.07, 1.58, "ornamental"]], "response_mode": "offline", "answer": "[{\"word\":\"ornamental\",\"start\":1.07,\"end\":1.58}]"} {"utterance_id": "1320-122617-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.53, 3.99, "balaam's"]], "response_mode": "offline", "answer": "[{\"word\":\"balaam's\",\"start\":3.53,\"end\":3.99}]"} {"utterance_id": "1320-122612-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0015.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.22, 2.54, "uncas"]], "response_mode": "offline", "answer": "[{\"word\":\"uncas\",\"start\":2.22,\"end\":2.54}]"} {"utterance_id": "6829-68771-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0014.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the full audio. After it finishes, find every instance of a family relation. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.09, 4.53, "cousin"]], "response_mode": "offline", "answer": "[{\"word\":\"cousin\",\"start\":4.09,\"end\":4.53}]"} {"utterance_id": "2830-3980-0070", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0070.flac", "task_type": "lexical_detection", "query": "sin", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"sin\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.86, 3.25, "sin"]], "response_mode": "offline", "answer": "[{\"word\":\"sin\",\"start\":2.86,\"end\":3.25}]"} {"utterance_id": "237-134500-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0012.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "When the utterance is complete, report how a number accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.5, 0.73, "one"], [8.63, 8.91, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.5,\"end\":0.73},{\"word\":\"two\",\"start\":8.63,\"end\":8.91}]"} {"utterance_id": "5142-36377-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.19, 1.01, "mister meadowcroft"]], "response_mode": "offline", "answer": "[{\"word\":\"mister meadowcroft\",\"start\":0.19,\"end\":1.01}]"} {"utterance_id": "1089-134691-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0023.flac", "task_type": "lexical_counting", "query": "pain", "prompt": "Listen to the entire clip, then report the count sequence for the word \"pain\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.74, 1.08, "one"], [2.61, 2.97, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.74,\"end\":1.08},{\"word\":\"two\",\"start\":2.61,\"end\":2.97}]"} {"utterance_id": "3570-5695-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0006.flac", "task_type": "lexical_detection", "query": "decency", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"decency\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[6.17, 6.65, "decency"]], "response_mode": "offline", "answer": "[{\"word\":\"decency\",\"start\":6.17,\"end\":6.65}]"} {"utterance_id": "1221-135766-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0012.flac", "task_type": "lexical_detection", "query": "irregularity", "prompt": "Once the audio has finished, identify every instance of the word \"irregularity\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.2, 7.17, "irregularity"]], "response_mode": "offline", "answer": "[{\"word\":\"irregularity\",\"start\":6.2,\"end\":7.17}]"} {"utterance_id": "8455-210777-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0030.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[2.44, 3.04, "captain"], [7.26, 7.79, "president"]], "response_mode": "offline", "answer": "[{\"word\":\"captain\",\"start\":2.44,\"end\":3.04},{\"word\":\"president\",\"start\":7.26,\"end\":7.79}]"} {"utterance_id": "672-122797-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0036.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 1.35, "humpy dumpy"]], "response_mode": "offline", "answer": "[{\"word\":\"humpy dumpy\",\"start\":0.52,\"end\":1.35}]"} {"utterance_id": "3570-5694-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0009.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.71, 12.41, "masters"]], "response_mode": "offline", "answer": "[{\"word\":\"masters\",\"start\":11.71,\"end\":12.41}]"} {"utterance_id": "4970-29095-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0004.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen to the full audio. After it finishes, find every instance of a money or currency term. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.45, 8.27, "compensation"]], "response_mode": "offline", "answer": "[{\"word\":\"compensation\",\"start\":7.45,\"end\":8.27}]"} {"utterance_id": "2830-3980-0060", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0060.flac", "task_type": "lexical_detection", "query": "loses", "prompt": "When the utterance is complete, report all instances of the word \"loses\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.57, 0.9, "loses"]], "response_mode": "offline", "answer": "[{\"word\":\"loses\",\"start\":0.57,\"end\":0.9}]"} {"utterance_id": "1221-135767-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0023.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[10.69, 11.02, "half"]], "response_mode": "offline", "answer": "[{\"word\":\"half\",\"start\":10.69,\"end\":11.02}]"} {"utterance_id": "7729-102255-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0016.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After listening to the whole recording, output every instance of an organization name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.24, 5.73, "militia"], [10.87, 11.46, "troops"]], "response_mode": "offline", "answer": "[{\"word\":\"militia\",\"start\":5.24,\"end\":5.73},{\"word\":\"troops\",\"start\":10.87,\"end\":11.46}]"} {"utterance_id": "121-123852-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123852/121-123852-0000.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.11, 7.68, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":7.11,\"end\":7.68}]"} {"utterance_id": "672-122797-0068", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0068.flac", "task_type": "lexical_detection", "query": "meant", "prompt": "When the utterance is complete, report all instances of the word \"meant\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.99, 3.5, "meant"]], "response_mode": "offline", "answer": "[{\"word\":\"meant\",\"start\":2.99,\"end\":3.5}]"} {"utterance_id": "3575-170457-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0024.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Once the audio has finished, identify every instance of a family relation. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[7.82, 8.41, "father's"]], "response_mode": "offline", "answer": "[{\"word\":\"father's\",\"start\":7.82,\"end\":8.41}]"} {"utterance_id": "7729-102255-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0004.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Listen carefully to the entire clip, then report every instance of a state or province name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.63, 1.04, "state"], [7.06, 7.49, "north"]], "response_mode": "offline", "answer": "[{\"word\":\"state\",\"start\":0.63,\"end\":1.04},{\"word\":\"north\",\"start\":7.06,\"end\":7.49}]"} {"utterance_id": "4992-41797-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0004.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.91, 6.33, "doctor"]], "response_mode": "offline", "answer": "[{\"word\":\"doctor\",\"start\":5.91,\"end\":6.33}]"} {"utterance_id": "8224-274384-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0004.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.32, 5.64, "ten"]], "response_mode": "offline", "answer": "[{\"word\":\"ten\",\"start\":5.32,\"end\":5.64}]"} {"utterance_id": "237-134500-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0016.flac", "task_type": "lexical_counting", "query": "know", "prompt": "Once the audio has finished, list each occurrence of the word \"know\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[0.79, 1.01, "one"], [2.04, 2.23, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.79,\"end\":1.01},{\"word\":\"two\",\"start\":2.04,\"end\":2.23}]"} {"utterance_id": "2961-961-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0021.flac", "task_type": "lexical_detection", "query": "wanted", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"wanted\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.96, 3.31, "wanted"]], "response_mode": "offline", "answer": "[{\"word\":\"wanted\",\"start\":2.96,\"end\":3.31}]"} {"utterance_id": "5683-32879-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0000.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Once the audio has finished, identify every instance of a city name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.44, 5.01, "brandon"]], "response_mode": "offline", "answer": "[{\"word\":\"brandon\",\"start\":4.44,\"end\":5.01}]"} {"utterance_id": "5683-32865-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.57, 1.85, "lake's"]], "response_mode": "offline", "answer": "[{\"word\":\"lake's\",\"start\":1.57,\"end\":1.85}]"} {"utterance_id": "1284-134647-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0007.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[10.87, 11.29, "bishops"], [13.1, 13.69, "primate"]], "response_mode": "offline", "answer": "[{\"word\":\"bishops\",\"start\":10.87,\"end\":11.29},{\"word\":\"primate\",\"start\":13.1,\"end\":13.69}]"} {"utterance_id": "2300-131720-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0018.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Listen carefully to the entire clip, then report every instance of a state or province name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.88, 8.6, "territory"], [10.63, 11.12, "districts"]], "response_mode": "offline", "answer": "[{\"word\":\"territory\",\"start\":7.88,\"end\":8.6},{\"word\":\"districts\",\"start\":10.63,\"end\":11.12}]"} {"utterance_id": "4446-2273-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0027.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.4, 1.98, "alexander"]], "response_mode": "offline", "answer": "[{\"word\":\"alexander\",\"start\":1.4,\"end\":1.98}]"} {"utterance_id": "3570-5696-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0007.flac", "task_type": "lexical_detection", "query": "implies", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"implies\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.9, 3.33, "implies"]], "response_mode": "offline", "answer": "[{\"word\":\"implies\",\"start\":2.9,\"end\":3.33}]"} {"utterance_id": "7021-85628-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0001.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.22, 10.94, "anders"]], "response_mode": "offline", "answer": "[{\"word\":\"anders\",\"start\":10.22,\"end\":10.94}]"} {"utterance_id": "2094-142345-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0031.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.89, 2.46, "aunt rachel"]], "response_mode": "offline", "answer": "[{\"word\":\"aunt rachel\",\"start\":1.89,\"end\":2.46}]"} {"utterance_id": "2830-3979-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0009.flac", "task_type": "lexical_detection", "query": "writings", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"writings\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.12, 2.49, "writings"]], "response_mode": "offline", "answer": "[{\"word\":\"writings\",\"start\":2.12,\"end\":2.49}]"} {"utterance_id": "61-70968-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0012.flac", "task_type": "lexical_detection", "query": "cries", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"cries\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.3, 0.68, "cries"]], "response_mode": "offline", "answer": "[{\"word\":\"cries\",\"start\":0.3,\"end\":0.68}]"} {"utterance_id": "121-121726-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0009.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Do not answer until the audio ends. Then list each instance of a money or currency term as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.9, 5.4, "dollars"]], "response_mode": "offline", "answer": "[{\"word\":\"dollars\",\"start\":4.9,\"end\":5.4}]"} {"utterance_id": "8230-279154-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0034.flac", "task_type": "lexical_counting", "query": "leads", "prompt": "Listen to the entire clip, then report the count sequence for the word \"leads\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.2, 4.46, "one"], [10.89, 11.21, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.2,\"end\":4.46},{\"word\":\"two\",\"start\":10.89,\"end\":11.21}]"} {"utterance_id": "3575-170457-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0042.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the full audio. After it finishes, find every instance of an ordinal number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.18, 3.41, "next"]], "response_mode": "offline", "answer": "[{\"word\":\"next\",\"start\":3.18,\"end\":3.41}]"} {"utterance_id": "3575-170457-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0041.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Once the audio has finished, identify every instance of a color name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[8.26, 8.56, "dark"]], "response_mode": "offline", "answer": "[{\"word\":\"dark\",\"start\":8.26,\"end\":8.56}]"} {"utterance_id": "237-134500-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0042.flac", "task_type": "lexical_detection", "query": "good", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"good\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.67, 0.96, "good"]], "response_mode": "offline", "answer": "[{\"word\":\"good\",\"start\":0.67,\"end\":0.96}]"} {"utterance_id": "1580-141084-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0005.flac", "task_type": "lexical_detection", "query": "important", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"important\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.05, 1.63, "important"]], "response_mode": "offline", "answer": "[{\"word\":\"important\",\"start\":1.05,\"end\":1.63}]"} {"utterance_id": "4992-41797-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0018.flac", "task_type": "lexical_detection", "query": "weighed", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"weighed\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[6.45, 6.78, "weighed"]], "response_mode": "offline", "answer": "[{\"word\":\"weighed\",\"start\":6.45,\"end\":6.78}]"} {"utterance_id": "1284-1180-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0008.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen carefully to the entire clip, then report every instance of a food name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.5, 7.86, "bread"]], "response_mode": "offline", "answer": "[{\"word\":\"bread\",\"start\":7.5,\"end\":7.86}]"} {"utterance_id": "1995-1837-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0012.flac", "task_type": "lexical_detection", "query": "splashed", "prompt": "Once the audio has finished, identify every instance of the word \"splashed\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.21, 0.79, "splashed"]], "response_mode": "offline", "answer": "[{\"word\":\"splashed\",\"start\":0.21,\"end\":0.79}]"} {"utterance_id": "2830-3979-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0001.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Once the audio has finished, list each occurrence of a person's name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[2.62, 2.93, "one"], [11.76, 12.16, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.62,\"end\":2.93},{\"word\":\"two\",\"start\":11.76,\"end\":12.16}]"} {"utterance_id": "3729-6852-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0032.flac", "task_type": "lexical_counting", "query": "snuff", "prompt": "Once the audio has finished, list each occurrence of the word \"snuff\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[5.41, 5.75, "one"], [10.49, 11.04, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.41,\"end\":5.75},{\"word\":\"two\",\"start\":10.49,\"end\":11.04}]"} {"utterance_id": "4446-2271-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0015.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.14, 1.53, "nerves"]], "response_mode": "offline", "answer": "[{\"word\":\"nerves\",\"start\":1.14,\"end\":1.53}]"} {"utterance_id": "1089-134686-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0011.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "After the audio ends, extract all occurrences of a time of day. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.14, 1.73, "mornings"]], "response_mode": "offline", "answer": "[{\"word\":\"mornings\",\"start\":1.14,\"end\":1.73}]"} {"utterance_id": "4446-2273-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0015.flac", "task_type": "lexical_detection", "query": "sorry", "prompt": "Once the audio has finished, identify every instance of the word \"sorry\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.64, 2.19, "sorry"]], "response_mode": "offline", "answer": "[{\"word\":\"sorry\",\"start\":1.64,\"end\":2.19}]"} {"utterance_id": "6829-68769-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0024.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 0.65, "will"], [3.0, 3.7, "missus rogers"]], "response_mode": "offline", "answer": "[{\"word\":\"will\",\"start\":0.42,\"end\":0.65},{\"word\":\"missus rogers\",\"start\":3.0,\"end\":3.7}]"} {"utterance_id": "237-134500-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0013.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "After listening to the whole recording, output timed count words for a plant name (one, two, three, ...) as JSON. If none, reply [].", "labels": [[1.1, 1.69, "one"], [4.94, 5.37, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.1,\"end\":1.69},{\"word\":\"two\",\"start\":4.94,\"end\":5.37}]"} {"utterance_id": "7176-88083-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0019.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.7, 3.25, "talons"], [5.21, 5.62, "head"]], "response_mode": "offline", "answer": "[{\"word\":\"talons\",\"start\":2.7,\"end\":3.25},{\"word\":\"head\",\"start\":5.21,\"end\":5.62}]"} {"utterance_id": "237-126133-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0001.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Listen to the full audio. After it finishes, find every instance of a weather term. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.2, 10.62, "wind"]], "response_mode": "offline", "answer": "[{\"word\":\"wind\",\"start\":10.2,\"end\":10.62}]"} {"utterance_id": "5639-40744-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0036.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.09, 8.56, "third"]], "response_mode": "offline", "answer": "[{\"word\":\"third\",\"start\":8.09,\"end\":8.56}]"} {"utterance_id": "7176-92135-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0000.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.26, 8.23, "actor managers"], [9.27, 9.65, "editors"]], "response_mode": "offline", "answer": "[{\"word\":\"actor managers\",\"start\":7.26,\"end\":8.23},{\"word\":\"editors\",\"start\":9.27,\"end\":9.65}]"} {"utterance_id": "7021-85628-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0007.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "Process the full audio offline. For every occurrence of a profession, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.45, 6.91, "one"], [11.9, 12.62, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.45,\"end\":6.91},{\"word\":\"two\",\"start\":11.9,\"end\":12.62}]"} {"utterance_id": "6930-76324-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0012.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen to the full audio. After it finishes, find every instance of a time of day. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.15, 1.69, "afternoon"]], "response_mode": "offline", "answer": "[{\"word\":\"afternoon\",\"start\":1.15,\"end\":1.69}]"} {"utterance_id": "4446-2273-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0001.flac", "task_type": "semantic_detection", "query": "city", "prompt": "When the utterance is complete, report all instances of a city name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.81, 2.36, "chelsea"]], "response_mode": "offline", "answer": "[{\"word\":\"chelsea\",\"start\":1.81,\"end\":2.36}]"} {"utterance_id": "1995-1826-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0002.flac", "task_type": "lexical_detection", "query": "college", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"college\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.17, 2.64, "college"]], "response_mode": "offline", "answer": "[{\"word\":\"college\",\"start\":2.17,\"end\":2.64}]"} {"utterance_id": "4446-2273-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0022.flac", "task_type": "lexical_detection", "query": "remembering", "prompt": "When the utterance is complete, report all instances of the word \"remembering\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.0, 1.48, "remembering"]], "response_mode": "offline", "answer": "[{\"word\":\"remembering\",\"start\":1.0,\"end\":1.48}]"} {"utterance_id": "260-123440-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0000.flac", "task_type": "lexical_detection", "query": "directions", "prompt": "Process the full audio offline. Find every instance of the word \"directions\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.97, 1.47, "directions"]], "response_mode": "offline", "answer": "[{\"word\":\"directions\",\"start\":0.97,\"end\":1.47}]"} {"utterance_id": "61-70968-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0046.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Do not answer until the audio ends. Then list each instance of a city name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.41, 0.86, "nottingham"]], "response_mode": "offline", "answer": "[{\"word\":\"nottingham\",\"start\":0.41,\"end\":0.86}]"} {"utterance_id": "4446-2275-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0030.flac", "task_type": "lexical_detection", "query": "simply", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"simply\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.86, 2.32, "simply"]], "response_mode": "offline", "answer": "[{\"word\":\"simply\",\"start\":1.86,\"end\":2.32}]"} {"utterance_id": "4970-29095-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0034.flac", "task_type": "lexical_detection", "query": "stupid", "prompt": "When the utterance is complete, report all instances of the word \"stupid\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.85, 2.51, "stupid"]], "response_mode": "offline", "answer": "[{\"word\":\"stupid\",\"start\":1.85,\"end\":2.51}]"} {"utterance_id": "7176-88083-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0006.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Once the audio has finished, identify every instance of an animal name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.37, 1.76, "trout"]], "response_mode": "offline", "answer": "[{\"word\":\"trout\",\"start\":1.37,\"end\":1.76}]"} {"utterance_id": "237-134500-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0002.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Listen to the full audio. After it finishes, find every instance of a weather term. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.76, 1.05, "wind"], [3.26, 3.71, "clouds"]], "response_mode": "offline", "answer": "[{\"word\":\"wind\",\"start\":0.76,\"end\":1.05},{\"word\":\"clouds\",\"start\":3.26,\"end\":3.71}]"} {"utterance_id": "8230-279154-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0040.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.16, 2.62, "several"]], "response_mode": "offline", "answer": "[{\"word\":\"several\",\"start\":2.16,\"end\":2.62}]"} {"utterance_id": "3570-5695-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0006.flac", "task_type": "semantic_detection", "query": "money", "prompt": "When the utterance is complete, report all instances of a money or currency term including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.54, 6.17, "pecuniary"]], "response_mode": "offline", "answer": "[{\"word\":\"pecuniary\",\"start\":5.54,\"end\":6.17}]"} {"utterance_id": "6829-68769-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0027.flac", "task_type": "lexical_detection", "query": "sign", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"sign\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.09, 1.35, "sign"]], "response_mode": "offline", "answer": "[{\"word\":\"sign\",\"start\":1.09,\"end\":1.35}]"} {"utterance_id": "1089-134691-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0009.flac", "task_type": "lexical_detection", "query": "gazing", "prompt": "Once the audio has finished, identify every instance of the word \"gazing\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.2, 4.67, "gazing"]], "response_mode": "offline", "answer": "[{\"word\":\"gazing\",\"start\":4.2,\"end\":4.67}]"} {"utterance_id": "5639-40744-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0012.flac", "task_type": "semantic_detection", "query": "color", "prompt": "You will hear a complete utterance. When it ends, list every instance of a color name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.94, 6.29, "colour"]], "response_mode": "offline", "answer": "[{\"word\":\"colour\",\"start\":5.94,\"end\":6.29}]"} {"utterance_id": "908-31957-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0004.flac", "task_type": "lexical_detection", "query": "blessing", "prompt": "Once the audio has finished, identify every instance of the word \"blessing\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.65, 3.3, "blessing"]], "response_mode": "offline", "answer": "[{\"word\":\"blessing\",\"start\":2.65,\"end\":3.3}]"} {"utterance_id": "260-123286-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0018.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.4, 2.92, "skeleton"]], "response_mode": "offline", "answer": "[{\"word\":\"skeleton\",\"start\":2.4,\"end\":2.92}]"} {"utterance_id": "2830-3979-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0006.flac", "task_type": "lexical_detection", "query": "luther's", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"luther's\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.45, 2.84, "luther's"]], "response_mode": "offline", "answer": "[{\"word\":\"luther's\",\"start\":2.45,\"end\":2.84}]"} {"utterance_id": "4077-13754-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0008.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Listen to the full audio. After it finishes, find every instance of a state or province name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[19.03, 19.48, "west"]], "response_mode": "offline", "answer": "[{\"word\":\"west\",\"start\":19.03,\"end\":19.48}]"} {"utterance_id": "2830-3979-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0011.flac", "task_type": "lexical_detection", "query": "teach", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"teach\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.15, 2.67, "teach"]], "response_mode": "offline", "answer": "[{\"word\":\"teach\",\"start\":2.15,\"end\":2.67}]"} {"utterance_id": "1089-134691-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0013.flac", "task_type": "lexical_counting", "query": "ourselves", "prompt": "Listen to the entire clip, then report the count sequence for the word \"ourselves\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.9, 9.68, "one"], [13.51, 14.16, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":8.9,\"end\":9.68},{\"word\":\"two\",\"start\":13.51,\"end\":14.16}]"} {"utterance_id": "61-70970-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0016.flac", "task_type": "lexical_detection", "query": "court", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"court\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.09, 3.39, "court"]], "response_mode": "offline", "answer": "[{\"word\":\"court\",\"start\":3.09,\"end\":3.39}]"} {"utterance_id": "7127-75946-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0010.flac", "task_type": "lexical_detection", "query": "majesty's", "prompt": "Once the audio has finished, identify every instance of the word \"majesty's\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.52, 1.15, "majesty's"]], "response_mode": "offline", "answer": "[{\"word\":\"majesty's\",\"start\":0.52,\"end\":1.15}]"} {"utterance_id": "1320-122612-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0000.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.31, 8.79, "hunter"]], "response_mode": "offline", "answer": "[{\"word\":\"hunter\",\"start\":8.31,\"end\":8.79}]"} {"utterance_id": "3570-5694-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0020.flac", "task_type": "lexical_detection", "query": "fighting", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"fighting\" with fields word, start, end (seconds). If none, reply [].", "labels": [[6.38, 6.79, "fighting"]], "response_mode": "offline", "answer": "[{\"word\":\"fighting\",\"start\":6.38,\"end\":6.79}]"} {"utterance_id": "8224-274381-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0004.flac", "task_type": "lexical_detection", "query": "levied", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"levied\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.8, 3.28, "levied"]], "response_mode": "offline", "answer": "[{\"word\":\"levied\",\"start\":2.8,\"end\":3.28}]"} {"utterance_id": "6829-68771-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0000.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Do not answer until the audio ends. Then list each instance of an organization name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.86, 2.92, "democratic committee"]], "response_mode": "offline", "answer": "[{\"word\":\"democratic committee\",\"start\":1.86,\"end\":2.92}]"} {"utterance_id": "6829-68769-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0021.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.0, 2.65, "tom gates"]], "response_mode": "offline", "answer": "[{\"word\":\"tom gates\",\"start\":2.0,\"end\":2.65}]"} {"utterance_id": "908-157963-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0025.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a family relation with fields word, start, end (seconds). If none, reply [].", "labels": [[7.28, 7.83, "mothers"]], "response_mode": "offline", "answer": "[{\"word\":\"mothers\",\"start\":7.28,\"end\":7.83}]"} {"utterance_id": "1580-141084-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0032.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.07, 1.6, "gilchrist"]], "response_mode": "offline", "answer": "[{\"word\":\"gilchrist\",\"start\":1.07,\"end\":1.6}]"} {"utterance_id": "3575-170457-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0033.flac", "task_type": "lexical_detection", "query": "comes", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"comes\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.65, 2.08, "comes"]], "response_mode": "offline", "answer": "[{\"word\":\"comes\",\"start\":1.65,\"end\":2.08}]"} {"utterance_id": "8455-210777-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0027.flac", "task_type": "lexical_detection", "query": "england", "prompt": "Once the audio has finished, identify every instance of the word \"england\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.3, 3.75, "england"]], "response_mode": "offline", "answer": "[{\"word\":\"england\",\"start\":3.3,\"end\":3.75}]"} {"utterance_id": "1188-133604-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0005.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.98, 4.91, "carpaccio's"], [7.22, 8.05, "saint george"]], "response_mode": "offline", "answer": "[{\"word\":\"carpaccio's\",\"start\":3.98,\"end\":4.91},{\"word\":\"saint george\",\"start\":7.22,\"end\":8.05}]"} {"utterance_id": "4507-16021-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0052.flac", "task_type": "lexical_detection", "query": "luminous", "prompt": "Once the audio has finished, identify every instance of the word \"luminous\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.59, 4.52, "luminous"]], "response_mode": "offline", "answer": "[{\"word\":\"luminous\",\"start\":3.59,\"end\":4.52}]"} {"utterance_id": "5142-36377-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0003.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.39, 0.89, "ambrose"]], "response_mode": "offline", "answer": "[{\"word\":\"ambrose\",\"start\":0.39,\"end\":0.89}]"} {"utterance_id": "672-122797-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0023.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.57, 9.19, "birds"]], "response_mode": "offline", "answer": "[{\"word\":\"birds\",\"start\":8.57,\"end\":9.19}]"} {"utterance_id": "6930-75918-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0013.flac", "task_type": "lexical_detection", "query": "terms", "prompt": "After the audio ends, extract all occurrences of the word \"terms\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.08, 1.52, "terms"]], "response_mode": "offline", "answer": "[{\"word\":\"terms\",\"start\":1.08,\"end\":1.52}]"} {"utterance_id": "5639-40744-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0029.flac", "task_type": "lexical_detection", "query": "learned", "prompt": "After listening to the whole recording, output every instance of the word \"learned\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.42, 1.79, "learned"]], "response_mode": "offline", "answer": "[{\"word\":\"learned\",\"start\":1.42,\"end\":1.79}]"} {"utterance_id": "5683-32879-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0020.flac", "task_type": "lexical_detection", "query": "threw", "prompt": "Process the full audio offline. Find every instance of the word \"threw\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.66, 1.07, "threw"]], "response_mode": "offline", "answer": "[{\"word\":\"threw\",\"start\":0.66,\"end\":1.07}]"} {"utterance_id": "2830-3980-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0036.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Once the audio has finished, identify every instance of an organization name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.85, 3.4, "church"]], "response_mode": "offline", "answer": "[{\"word\":\"church\",\"start\":2.85,\"end\":3.4}]"} {"utterance_id": "1188-133604-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0026.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a plant name with word/start/end. If none, reply [].", "labels": [[6.31, 6.74, "one"], [8.9, 9.43, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.31,\"end\":6.74},{\"word\":\"two\",\"start\":8.9,\"end\":9.43}]"} {"utterance_id": "8555-284449-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.06, 1.53, "rosalie"]], "response_mode": "offline", "answer": "[{\"word\":\"rosalie\",\"start\":1.06,\"end\":1.53}]"} {"utterance_id": "61-70968-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0014.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[2.92, 3.3, "squire"]], "response_mode": "offline", "answer": "[{\"word\":\"squire\",\"start\":2.92,\"end\":3.3}]"} {"utterance_id": "4992-41797-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0019.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.54, 1.16, "olive's"], [2.62, 3.08, "nancy's"]], "response_mode": "offline", "answer": "[{\"word\":\"olive's\",\"start\":0.54,\"end\":1.16},{\"word\":\"nancy's\",\"start\":2.62,\"end\":3.08}]"} {"utterance_id": "8555-284447-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0000.flac", "task_type": "lexical_detection", "query": "courtyard", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"courtyard\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.99, 2.68, "courtyard"]], "response_mode": "offline", "answer": "[{\"word\":\"courtyard\",\"start\":1.99,\"end\":2.68}]"} {"utterance_id": "2300-131720-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0036.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.03, 4.11, "eighteen ninety nine"]], "response_mode": "offline", "answer": "[{\"word\":\"eighteen ninety nine\",\"start\":3.03,\"end\":4.11}]"} {"utterance_id": "1995-1826-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0021.flac", "task_type": "lexical_detection", "query": "literally", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"literally\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.64, 5.14, "literally"]], "response_mode": "offline", "answer": "[{\"word\":\"literally\",\"start\":4.64,\"end\":5.14}]"} {"utterance_id": "3570-5695-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0005.flac", "task_type": "semantic_detection", "query": "money", "prompt": "After listening to the whole recording, output every instance of a money or currency term with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[18.36, 18.95, "pecuniary"]], "response_mode": "offline", "answer": "[{\"word\":\"pecuniary\",\"start\":18.36,\"end\":18.95}]"} {"utterance_id": "8224-274384-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0008.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an ordinal number with fields word, start, end (seconds). If none, reply [].", "labels": [[11.49, 11.93, "former"]], "response_mode": "offline", "answer": "[{\"word\":\"former\",\"start\":11.49,\"end\":11.93}]"} {"utterance_id": "6930-81414-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0024.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.35, 0.67, "tongue"]], "response_mode": "offline", "answer": "[{\"word\":\"tongue\",\"start\":0.35,\"end\":0.67}]"} {"utterance_id": "1284-134647-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0007.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.19, 10.87, "four hundred"]], "response_mode": "offline", "answer": "[{\"word\":\"four hundred\",\"start\":10.19,\"end\":10.87}]"} {"utterance_id": "5142-36377-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0014.flac", "task_type": "lexical_counting", "query": "accent", "prompt": "After the audio ends, tally every instance of the word \"accent\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[18.6, 19.24, "one"], [23.41, 23.91, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":18.6,\"end\":19.24},{\"word\":\"two\",\"start\":23.41,\"end\":23.91}]"} {"utterance_id": "2830-3980-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0016.flac", "task_type": "lexical_detection", "query": "best", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"best\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.37, 1.71, "best"]], "response_mode": "offline", "answer": "[{\"word\":\"best\",\"start\":1.37,\"end\":1.71}]"} {"utterance_id": "7176-92135-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0039.flac", "task_type": "lexical_detection", "query": "butler", "prompt": "After listening to the whole recording, output every instance of the word \"butler\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.87, 2.32, "butler"]], "response_mode": "offline", "answer": "[{\"word\":\"butler\",\"start\":1.87,\"end\":2.32}]"} {"utterance_id": "7021-79730-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0009.flac", "task_type": "lexical_detection", "query": "skill", "prompt": "Process the full audio offline. Find every instance of the word \"skill\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.7, 5.32, "skill"]], "response_mode": "offline", "answer": "[{\"word\":\"skill\",\"start\":4.7,\"end\":5.32}]"} {"utterance_id": "1188-133604-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0005.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Listen to the full audio. Count every occurrence of a body part. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.84, 1.03, "one"], [3.02, 3.36, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.84,\"end\":1.03},{\"word\":\"two\",\"start\":3.02,\"end\":3.36}]"} {"utterance_id": "4077-13754-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0009.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen carefully to the entire clip, then report every instance of an organization name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.38, 3.31, "latter day saints"]], "response_mode": "offline", "answer": "[{\"word\":\"latter day saints\",\"start\":2.38,\"end\":3.31}]"} {"utterance_id": "4446-2271-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0020.flac", "task_type": "lexical_detection", "query": "reflected", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"reflected\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.89, 1.43, "reflected"]], "response_mode": "offline", "answer": "[{\"word\":\"reflected\",\"start\":0.89,\"end\":1.43}]"} {"utterance_id": "1089-134686-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0021.flac", "task_type": "lexical_detection", "query": "pour", "prompt": "Once the audio has finished, identify every instance of the word \"pour\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.33, 2.57, "pour"]], "response_mode": "offline", "answer": "[{\"word\":\"pour\",\"start\":2.33,\"end\":2.57}]"} {"utterance_id": "3575-170457-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0025.flac", "task_type": "lexical_detection", "query": "renewed", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"renewed\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.1, 3.46, "renewed"]], "response_mode": "offline", "answer": "[{\"word\":\"renewed\",\"start\":3.1,\"end\":3.46}]"} {"utterance_id": "260-123288-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0010.flac", "task_type": "semantic_counting", "query": "weather", "prompt": "Process the full audio offline. For every occurrence of a weather term, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.46, 4.62, "one"], [7.51, 7.95, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.46,\"end\":4.62},{\"word\":\"two\",\"start\":7.51,\"end\":7.95}]"} {"utterance_id": "7729-102255-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0003.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen carefully to the entire clip, then report every instance of an organization name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.04, 6.84, "factions"], [12.41, 13.03, "reserve"]], "response_mode": "offline", "answer": "[{\"word\":\"factions\",\"start\":6.04,\"end\":6.84},{\"word\":\"reserve\",\"start\":12.41,\"end\":13.03}]"} {"utterance_id": "4507-16021-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0015.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "Listen carefully to the entire clip, then report every instance of a disease name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.33, 1.95, "malady"]], "response_mode": "offline", "answer": "[{\"word\":\"malady\",\"start\":1.33,\"end\":1.95}]"} {"utterance_id": "61-70968-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0032.flac", "task_type": "lexical_detection", "query": "cloth", "prompt": "Process the full audio offline. Find every instance of the word \"cloth\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.81, 2.3, "cloth"]], "response_mode": "offline", "answer": "[{\"word\":\"cloth\",\"start\":1.81,\"end\":2.3}]"} {"utterance_id": "6829-68769-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0011.flac", "task_type": "lexical_detection", "query": "quietly", "prompt": "When the utterance is complete, report all instances of the word \"quietly\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.77, 4.36, "quietly"]], "response_mode": "offline", "answer": "[{\"word\":\"quietly\",\"start\":3.77,\"end\":4.36}]"} {"utterance_id": "5683-32879-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0014.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a plant name with fields word, start, end (seconds). If none, reply [].", "labels": [[5.47, 6.01, "flower"]], "response_mode": "offline", "answer": "[{\"word\":\"flower\",\"start\":5.47,\"end\":6.01}]"} {"utterance_id": "5105-28240-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0022.flac", "task_type": "lexical_detection", "query": "schooner", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"schooner\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.25, 3.7, "schooner"]], "response_mode": "offline", "answer": "[{\"word\":\"schooner\",\"start\":3.25,\"end\":3.7}]"} {"utterance_id": "7176-88083-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0002.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.55, 0.84, "feet"], [2.68, 3.07, "beak"]], "response_mode": "offline", "answer": "[{\"word\":\"feet\",\"start\":0.55,\"end\":0.84},{\"word\":\"beak\",\"start\":2.68,\"end\":3.07}]"} {"utterance_id": "1580-141084-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0004.flac", "task_type": "lexical_detection", "query": "realize", "prompt": "When the utterance is complete, report all instances of the word \"realize\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.28, 1.77, "realize"]], "response_mode": "offline", "answer": "[{\"word\":\"realize\",\"start\":1.28,\"end\":1.77}]"} {"utterance_id": "6829-68771-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0032.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Listen to the entire clip, then report the count sequence for a body part as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.89, 1.38, "one"], [4.68, 5.1, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.89,\"end\":1.38},{\"word\":\"two\",\"start\":4.68,\"end\":5.1}]"} {"utterance_id": "4077-13751-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0006.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Process the full audio offline. Find every instance of a time of day and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.9, 4.29, "evening"]], "response_mode": "offline", "answer": "[{\"word\":\"evening\",\"start\":3.9,\"end\":4.29}]"} {"utterance_id": "8555-284447-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.71, 3.37, "cap'n bill"]], "response_mode": "offline", "answer": "[{\"word\":\"cap'n bill\",\"start\":2.71,\"end\":3.37}]"} {"utterance_id": "1089-134686-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0034.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.66, 1.0, "rector"]], "response_mode": "offline", "answer": "[{\"word\":\"rector\",\"start\":0.66,\"end\":1.0}]"} {"utterance_id": "2830-3980-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0021.flac", "task_type": "lexical_detection", "query": "father", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"father\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.86, 1.24, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":0.86,\"end\":1.24}]"} {"utterance_id": "2830-3980-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0025.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.48, 1.82, "christ"]], "response_mode": "offline", "answer": "[{\"word\":\"christ\",\"start\":1.48,\"end\":1.82}]"} {"utterance_id": "6829-68771-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0024.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[1.57, 1.86, "maid"]], "response_mode": "offline", "answer": "[{\"word\":\"maid\",\"start\":1.57,\"end\":1.86}]"} {"utterance_id": "1995-1826-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0008.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the full audio. After it finishes, find every instance of a plant name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.6, 2.0, "cotton"]], "response_mode": "offline", "answer": "[{\"word\":\"cotton\",\"start\":1.6,\"end\":2.0}]"} {"utterance_id": "1995-1837-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0022.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a color name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.51, 2.81, "white"]], "response_mode": "offline", "answer": "[{\"word\":\"white\",\"start\":2.51,\"end\":2.81}]"} {"utterance_id": "1188-133604-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0036.flac", "task_type": "lexical_detection", "query": "subjects", "prompt": "Once the audio has finished, identify every instance of the word \"subjects\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.27, 3.04, "subjects"]], "response_mode": "offline", "answer": "[{\"word\":\"subjects\",\"start\":2.27,\"end\":3.04}]"} {"utterance_id": "672-122797-0062", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0062.flac", "task_type": "lexical_detection", "query": "tree", "prompt": "After listening to the whole recording, output every instance of the word \"tree\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.95, 2.42, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":1.95,\"end\":2.42}]"} {"utterance_id": "672-122797-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0059.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.38, 1.87, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.38,\"end\":1.87}]"} {"utterance_id": "5683-32866-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0028.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.1, 1.75, "trees"]], "response_mode": "offline", "answer": "[{\"word\":\"trees\",\"start\":1.1,\"end\":1.75}]"} {"utterance_id": "1320-122612-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0008.flac", "task_type": "lexical_detection", "query": "unexpected", "prompt": "When the utterance is complete, report all instances of the word \"unexpected\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.28, 2.9, "unexpected"]], "response_mode": "offline", "answer": "[{\"word\":\"unexpected\",\"start\":2.28,\"end\":2.9}]"} {"utterance_id": "237-134493-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0002.flac", "task_type": "lexical_detection", "query": "graveyard", "prompt": "Once the audio has finished, identify every instance of the word \"graveyard\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.23, 1.96, "graveyard"]], "response_mode": "offline", "answer": "[{\"word\":\"graveyard\",\"start\":1.23,\"end\":1.96}]"} {"utterance_id": "7729-102255-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0015.flac", "task_type": "semantic_counting", "query": "city", "prompt": "After listening to the whole recording, output timed count words for a city name (one, two, three, ...) as JSON. If none, reply [].", "labels": [[6.07, 6.68, "one"], [13.89, 14.82, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.07,\"end\":6.68},{\"word\":\"two\",\"start\":13.89,\"end\":14.82}]"} {"utterance_id": "7729-102255-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0043.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.57, 4.53, "governor shannon"]], "response_mode": "offline", "answer": "[{\"word\":\"governor shannon\",\"start\":3.57,\"end\":4.53}]"} {"utterance_id": "1580-141084-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0042.flac", "task_type": "lexical_detection", "query": "impulse", "prompt": "After listening to the whole recording, output every instance of the word \"impulse\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.84, 1.34, "impulse"]], "response_mode": "offline", "answer": "[{\"word\":\"impulse\",\"start\":0.84,\"end\":1.34}]"} {"utterance_id": "2094-142345-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0031.flac", "task_type": "lexical_detection", "query": "fond", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"fond\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.1, 1.32, "fond"]], "response_mode": "offline", "answer": "[{\"word\":\"fond\",\"start\":1.1,\"end\":1.32}]"} {"utterance_id": "1284-1180-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0006.flac", "task_type": "lexical_counting", "query": "country", "prompt": "Listen to the full audio. Count every occurrence of the word \"country\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.77, "one"], [3.92, 4.29, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.35,\"end\":2.77},{\"word\":\"two\",\"start\":3.92,\"end\":4.29}]"} {"utterance_id": "672-122797-0060", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0060.flac", "task_type": "lexical_detection", "query": "stupid", "prompt": "After listening to the whole recording, output every instance of the word \"stupid\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.67, 2.29, "stupid"]], "response_mode": "offline", "answer": "[{\"word\":\"stupid\",\"start\":1.67,\"end\":2.29}]"} {"utterance_id": "7729-102255-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0015.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.0, 0.5, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.0,\"end\":0.5}]"} {"utterance_id": "7729-102255-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0032.flac", "task_type": "lexical_counting", "query": "free", "prompt": "When the utterance is complete, report how the word \"free\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.54, 0.75, "one"], [17.95, 18.28, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.54,\"end\":0.75},{\"word\":\"two\",\"start\":17.95,\"end\":18.28}]"} {"utterance_id": "1284-1180-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0027.flac", "task_type": "lexical_detection", "query": "task", "prompt": "Once the audio has finished, identify every instance of the word \"task\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.69, 1.11, "task"]], "response_mode": "offline", "answer": "[{\"word\":\"task\",\"start\":0.69,\"end\":1.11}]"} {"utterance_id": "4507-16021-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0047.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[28.19, 29.56, "vertebral column"]], "response_mode": "offline", "answer": "[{\"word\":\"vertebral column\",\"start\":28.19,\"end\":29.56}]"} {"utterance_id": "3729-6852-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0012.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.54, 2.35, "esprit"]], "response_mode": "offline", "answer": "[{\"word\":\"esprit\",\"start\":1.54,\"end\":2.35}]"} {"utterance_id": "908-157963-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0001.flac", "task_type": "lexical_detection", "query": "spring", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"spring\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.05, 2.68, "spring"]], "response_mode": "offline", "answer": "[{\"word\":\"spring\",\"start\":2.05,\"end\":2.68}]"} {"utterance_id": "4970-29093-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0018.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.55, 0.9, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":0.55,\"end\":0.9}]"} {"utterance_id": "260-123286-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0027.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.2, 1.54, "eye"], [6.46, 6.96, "head"]], "response_mode": "offline", "answer": "[{\"word\":\"eye\",\"start\":1.2,\"end\":1.54},{\"word\":\"head\",\"start\":6.46,\"end\":6.96}]"} {"utterance_id": "4446-2275-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0029.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.89, 1.1, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.89,\"end\":1.1}]"} {"utterance_id": "1188-133604-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0043.flac", "task_type": "lexical_detection", "query": "boy's", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"boy's\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.69, "boy's"]], "response_mode": "offline", "answer": "[{\"word\":\"boy's\",\"start\":2.35,\"end\":2.69}]"} {"utterance_id": "2830-3980-0058", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0058.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 0.95, "mohammed"], [1.96, 2.42, "christ"]], "response_mode": "offline", "answer": "[{\"word\":\"mohammed\",\"start\":0.51,\"end\":0.95},{\"word\":\"christ\",\"start\":1.96,\"end\":2.42}]"} {"utterance_id": "61-70970-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0002.flac", "task_type": "lexical_detection", "query": "robin", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"robin\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.09, 1.42, "robin"]], "response_mode": "offline", "answer": "[{\"word\":\"robin\",\"start\":1.09,\"end\":1.42}]"} {"utterance_id": "7176-92135-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0015.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.17, 2.52, "ophelia"]], "response_mode": "offline", "answer": "[{\"word\":\"ophelia\",\"start\":2.17,\"end\":2.52}]"} {"utterance_id": "8224-274381-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0012.flac", "task_type": "lexical_counting", "query": "horse", "prompt": "Once the audio has finished, list each occurrence of the word \"horse\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[4.3, 4.75, "one"], [7.65, 8.02, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.3,\"end\":4.75},{\"word\":\"two\",\"start\":7.65,\"end\":8.02}]"} {"utterance_id": "2830-3980-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0053.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the full audio. Count every occurrence of a person's name. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.62, 1.86, "one"], [3.35, 3.78, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.62,\"end\":1.86},{\"word\":\"two\",\"start\":3.35,\"end\":3.78}]"} {"utterance_id": "1221-135766-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.85, 5.29, "pearl"]], "response_mode": "offline", "answer": "[{\"word\":\"pearl\",\"start\":4.85,\"end\":5.29}]"} {"utterance_id": "260-123440-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0019.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 0.76, "alice"]], "response_mode": "offline", "answer": "[{\"word\":\"alice\",\"start\":0.42,\"end\":0.76}]"} {"utterance_id": "121-127105-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0024.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Do not answer until the audio ends. Then list each instance of a time of day as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[13.28, 13.6, "night"]], "response_mode": "offline", "answer": "[{\"word\":\"night\",\"start\":13.28,\"end\":13.6}]"} {"utterance_id": "61-70970-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0032.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.83, 1.21, "robin"]], "response_mode": "offline", "answer": "[{\"word\":\"robin\",\"start\":0.83,\"end\":1.21}]"} {"utterance_id": "3729-6852-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0006.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.17, 14.81, "sisters"]], "response_mode": "offline", "answer": "[{\"word\":\"sisters\",\"start\":14.17,\"end\":14.81}]"} {"utterance_id": "4507-16021-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0023.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.91, 2.05, "manufacturer"], [9.5, 10.61, "manufacturer"]], "response_mode": "offline", "answer": "[{\"word\":\"manufacturer\",\"start\":0.91,\"end\":2.05},{\"word\":\"manufacturer\",\"start\":9.5,\"end\":10.61}]"} {"utterance_id": "3575-170457-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0048.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen carefully to the entire clip, then report every instance of an ordinal number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.1, 1.44, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":1.1,\"end\":1.44}]"} {"utterance_id": "5105-28241-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0008.flac", "task_type": "lexical_detection", "query": "orbit", "prompt": "Process the full audio offline. Find every instance of the word \"orbit\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.7, 3.2, "orbit"]], "response_mode": "offline", "answer": "[{\"word\":\"orbit\",\"start\":2.7,\"end\":3.2}]"} {"utterance_id": "1580-141083-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0026.flac", "task_type": "lexical_detection", "query": "entered", "prompt": "Once the audio has finished, identify every instance of the word \"entered\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.03, 3.35, "entered"]], "response_mode": "offline", "answer": "[{\"word\":\"entered\",\"start\":3.03,\"end\":3.35}]"} {"utterance_id": "8555-292519-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0001.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[6.51, 6.79, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.51,\"end\":6.79}]"} {"utterance_id": "1089-134686-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0005.flac", "task_type": "lexical_detection", "query": "fragment", "prompt": "Once the audio has finished, identify every instance of the word \"fragment\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.74, 5.28, "fragment"]], "response_mode": "offline", "answer": "[{\"word\":\"fragment\",\"start\":4.74,\"end\":5.28}]"} {"utterance_id": "5683-32866-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0009.flac", "task_type": "lexical_detection", "query": "obscure", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"obscure\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.87, 6.47, "obscure"]], "response_mode": "offline", "answer": "[{\"word\":\"obscure\",\"start\":5.87,\"end\":6.47}]"} {"utterance_id": "7021-85628-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0015.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.53, 8.92, "neck"]], "response_mode": "offline", "answer": "[{\"word\":\"neck\",\"start\":8.53,\"end\":8.92}]"} {"utterance_id": "8555-284447-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0012.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[0.56, 0.99, "captain"]], "response_mode": "offline", "answer": "[{\"word\":\"captain\",\"start\":0.56,\"end\":0.99}]"} {"utterance_id": "1284-1181-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0002.flac", "task_type": "lexical_detection", "query": "curious", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"curious\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.4, 2.95, "curious"]], "response_mode": "offline", "answer": "[{\"word\":\"curious\",\"start\":2.4,\"end\":2.95}]"} {"utterance_id": "5639-40744-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0024.flac", "task_type": "lexical_detection", "query": "relation", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"relation\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.91, 4.49, "relation"]], "response_mode": "offline", "answer": "[{\"word\":\"relation\",\"start\":3.91,\"end\":4.49}]"} {"utterance_id": "1320-122617-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0027.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.6, 2.96, "scout"]], "response_mode": "offline", "answer": "[{\"word\":\"scout\",\"start\":2.6,\"end\":2.96}]"} {"utterance_id": "61-70970-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0012.flac", "task_type": "lexical_detection", "query": "tricks", "prompt": "After listening to the whole recording, output every instance of the word \"tricks\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.51, 1.83, "tricks"]], "response_mode": "offline", "answer": "[{\"word\":\"tricks\",\"start\":1.51,\"end\":1.83}]"} {"utterance_id": "1221-135767-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0021.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.17, 8.63, "grass"], [11.93, 12.44, "shrubbery"]], "response_mode": "offline", "answer": "[{\"word\":\"grass\",\"start\":8.17,\"end\":8.63},{\"word\":\"shrubbery\",\"start\":11.93,\"end\":12.44}]"} {"utterance_id": "1221-135767-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0009.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.69, 5.07, "foot"], [6.12, 6.52, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"foot\",\"start\":4.69,\"end\":5.07},{\"word\":\"hand\",\"start\":6.12,\"end\":6.52}]"} {"utterance_id": "672-122797-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0059.flac", "task_type": "lexical_detection", "query": "answered", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"answered\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.3, 2.72, "answered"]], "response_mode": "offline", "answer": "[{\"word\":\"answered\",\"start\":2.3,\"end\":2.72}]"} {"utterance_id": "7127-75947-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0022.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.15, 10.8, "montalais"]], "response_mode": "offline", "answer": "[{\"word\":\"montalais\",\"start\":10.15,\"end\":10.8}]"} {"utterance_id": "4992-41797-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0015.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.21, 0.67, "cyril"]], "response_mode": "offline", "answer": "[{\"word\":\"cyril\",\"start\":0.21,\"end\":0.67}]"} {"utterance_id": "5639-40744-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0029.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a body part: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[2.29, 2.64, "one"], [3.87, 4.23, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.29,\"end\":2.64},{\"word\":\"two\",\"start\":3.87,\"end\":4.23}]"} {"utterance_id": "6829-68769-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0050.flac", "task_type": "lexical_detection", "query": "silence", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"silence\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.42, 1.96, "silence"]], "response_mode": "offline", "answer": "[{\"word\":\"silence\",\"start\":1.42,\"end\":1.96}]"} {"utterance_id": "4970-29093-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0003.flac", "task_type": "lexical_detection", "query": "modest", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"modest\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.61, 0.97, "modest"]], "response_mode": "offline", "answer": "[{\"word\":\"modest\",\"start\":0.61,\"end\":0.97}]"} {"utterance_id": "4077-13754-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0005.flac", "task_type": "semantic_detection", "query": "country", "prompt": "When the utterance is complete, report all instances of a country name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[16.46, 16.94, "country's"]], "response_mode": "offline", "answer": "[{\"word\":\"country's\",\"start\":16.46,\"end\":16.94}]"} {"utterance_id": "4446-2275-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0010.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.31, 2.59, "hands"]], "response_mode": "offline", "answer": "[{\"word\":\"hands\",\"start\":2.31,\"end\":2.59}]"} {"utterance_id": "908-157963-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0029.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Process the full audio offline. Find every instance of a food name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.85, 2.27, "honey"]], "response_mode": "offline", "answer": "[{\"word\":\"honey\",\"start\":1.85,\"end\":2.27}]"} {"utterance_id": "5683-32866-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0013.flac", "task_type": "lexical_counting", "query": "shadows", "prompt": "Once the audio has finished, list each occurrence of the word \"shadows\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[2.12, 2.6, "one"], [6.31, 6.87, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.12,\"end\":2.6},{\"word\":\"two\",\"start\":6.31,\"end\":6.87}]"} {"utterance_id": "5639-40744-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0015.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.26, 1.93, "rodolfo"]], "response_mode": "offline", "answer": "[{\"word\":\"rodolfo\",\"start\":1.26,\"end\":1.93}]"} {"utterance_id": "1995-1826-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.36, 1.07, "john taylor"]], "response_mode": "offline", "answer": "[{\"word\":\"john taylor\",\"start\":0.36,\"end\":1.07}]"} {"utterance_id": "260-123440-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0002.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an animal name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.12, 1.49, "rabbit"]], "response_mode": "offline", "answer": "[{\"word\":\"rabbit\",\"start\":1.12,\"end\":1.49}]"} {"utterance_id": "7021-85628-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0002.flac", "task_type": "lexical_detection", "query": "jack", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"jack\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.33, 5.7, "jack"]], "response_mode": "offline", "answer": "[{\"word\":\"jack\",\"start\":5.33,\"end\":5.7}]"} {"utterance_id": "61-70970-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.09, 1.42, "robin"]], "response_mode": "offline", "answer": "[{\"word\":\"robin\",\"start\":1.09,\"end\":1.42}]"} {"utterance_id": "1580-141083-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0012.flac", "task_type": "semantic_detection", "query": "food", "prompt": "When the utterance is complete, report all instances of a food name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.44, 3.81, "dough"]], "response_mode": "offline", "answer": "[{\"word\":\"dough\",\"start\":3.44,\"end\":3.81}]"} {"utterance_id": "4970-29095-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0032.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.6, 0.96, "philip"]], "response_mode": "offline", "answer": "[{\"word\":\"philip\",\"start\":0.6,\"end\":0.96}]"} {"utterance_id": "2300-131720-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0011.flac", "task_type": "lexical_detection", "query": "group", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"group\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.22, 4.53, "group"]], "response_mode": "offline", "answer": "[{\"word\":\"group\",\"start\":4.22,\"end\":4.53}]"} {"utterance_id": "1320-122617-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0040.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Do not answer until the audio ends. Then list each instance of an ordinal number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.34, 2.82, "next"]], "response_mode": "offline", "answer": "[{\"word\":\"next\",\"start\":2.34,\"end\":2.82}]"} {"utterance_id": "5683-32879-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0021.flac", "task_type": "lexical_detection", "query": "dorcas", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"dorcas\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.39, 0.98, "dorcas"]], "response_mode": "offline", "answer": "[{\"word\":\"dorcas\",\"start\":0.39,\"end\":0.98}]"} {"utterance_id": "1188-133604-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0037.flac", "task_type": "lexical_detection", "query": "traces", "prompt": "After the audio ends, extract all occurrences of the word \"traces\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.98, 2.47, "traces"]], "response_mode": "offline", "answer": "[{\"word\":\"traces\",\"start\":1.98,\"end\":2.47}]"} {"utterance_id": "8455-210777-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0023.flac", "task_type": "semantic_detection", "query": "food", "prompt": "After the audio ends, extract all occurrences of a food name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.58, 3.01, "dinner"]], "response_mode": "offline", "answer": "[{\"word\":\"dinner\",\"start\":2.58,\"end\":3.01}]"} {"utterance_id": "1089-134686-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0024.flac", "task_type": "lexical_detection", "query": "consecrated", "prompt": "After listening to the whole recording, output every instance of the word \"consecrated\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.28, 6.06, "consecrated"]], "response_mode": "offline", "answer": "[{\"word\":\"consecrated\",\"start\":5.28,\"end\":6.06}]"} {"utterance_id": "4077-13751-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0011.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.65, 1.18, "thousands"]], "response_mode": "offline", "answer": "[{\"word\":\"thousands\",\"start\":0.65,\"end\":1.18}]"} {"utterance_id": "8555-284447-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0019.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.28, 4.56, "king"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":4.28,\"end\":4.56}]"} {"utterance_id": "5142-33396-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0003.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.4, 2.7, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":2.4,\"end\":2.7}]"} {"utterance_id": "7176-92135-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0003.flac", "task_type": "lexical_counting", "query": "play", "prompt": "Process the full audio offline. For every occurrence of the word \"play\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.39, 0.66, "one"], [2.12, 2.45, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.39,\"end\":0.66},{\"word\":\"two\",\"start\":2.12,\"end\":2.45}]"} {"utterance_id": "672-122797-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0026.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.37, 12.89, "children"]], "response_mode": "offline", "answer": "[{\"word\":\"children\",\"start\":12.37,\"end\":12.89}]"} {"utterance_id": "260-123286-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0003.flac", "task_type": "lexical_counting", "query": "anxious", "prompt": "Once the audio has finished, list each occurrence of the word \"anxious\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[0.95, 1.47, "one"], [6.3, 7.09, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.95,\"end\":1.47},{\"word\":\"two\",\"start\":6.3,\"end\":7.09}]"} {"utterance_id": "5142-33396-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0005.flac", "task_type": "lexical_detection", "query": "stayed", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"stayed\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.46, 0.77, "stayed"]], "response_mode": "offline", "answer": "[{\"word\":\"stayed\",\"start\":0.46,\"end\":0.77}]"} {"utterance_id": "4507-16021-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0014.flac", "task_type": "lexical_detection", "query": "excluded", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"excluded\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.39, 5.11, "excluded"]], "response_mode": "offline", "answer": "[{\"word\":\"excluded\",\"start\":4.39,\"end\":5.11}]"} {"utterance_id": "121-127105-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0027.flac", "task_type": "semantic_detection", "query": "city", "prompt": "After the audio ends, extract all occurrences of a city name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.18, 1.5, "town"]], "response_mode": "offline", "answer": "[{\"word\":\"town\",\"start\":1.18,\"end\":1.5}]"} {"utterance_id": "1089-134686-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0009.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Once the audio has finished, identify every instance of a money or currency term. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.79, 2.22, "alms"]], "response_mode": "offline", "answer": "[{\"word\":\"alms\",\"start\":1.79,\"end\":2.22}]"} {"utterance_id": "1089-134686-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.93, 2.37, "nelly"]], "response_mode": "offline", "answer": "[{\"word\":\"nelly\",\"start\":1.93,\"end\":2.37}]"} {"utterance_id": "8455-210777-0049", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0049.flac", "task_type": "lexical_detection", "query": "officer", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"officer\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.04, 3.7, "officer"]], "response_mode": "offline", "answer": "[{\"word\":\"officer\",\"start\":3.04,\"end\":3.7}]"} {"utterance_id": "8555-284447-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0016.flac", "task_type": "lexical_detection", "query": "glorious", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"glorious\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.0, 1.93, "glorious"]], "response_mode": "offline", "answer": "[{\"word\":\"glorious\",\"start\":1.0,\"end\":1.93}]"} {"utterance_id": "6930-76324-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0004.flac", "task_type": "lexical_detection", "query": "joyce", "prompt": "Process the full audio offline. Find every instance of the word \"joyce\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.41, 0.79, "joyce"]], "response_mode": "offline", "answer": "[{\"word\":\"joyce\",\"start\":0.41,\"end\":0.79}]"} {"utterance_id": "4992-41806-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0012.flac", "task_type": "lexical_detection", "query": "bearing", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"bearing\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.46, 2.88, "bearing"]], "response_mode": "offline", "answer": "[{\"word\":\"bearing\",\"start\":2.46,\"end\":2.88}]"} {"utterance_id": "3575-170457-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0004.flac", "task_type": "lexical_detection", "query": "dispute", "prompt": "After the audio ends, extract all occurrences of the word \"dispute\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.93, 1.36, "dispute"]], "response_mode": "offline", "answer": "[{\"word\":\"dispute\",\"start\":0.93,\"end\":1.36}]"} {"utterance_id": "8230-279154-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0042.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.47, 13.66, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":13.47,\"end\":13.66}]"} {"utterance_id": "2300-131720-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0006.flac", "task_type": "lexical_detection", "query": "believing", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"believing\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.27, 2.71, "believing"]], "response_mode": "offline", "answer": "[{\"word\":\"believing\",\"start\":2.27,\"end\":2.71}]"} {"utterance_id": "7127-75946-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0003.flac", "task_type": "lexical_detection", "query": "gentlemen", "prompt": "Process the full audio offline. Find every instance of the word \"gentlemen\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.17, 0.63, "gentlemen"]], "response_mode": "offline", "answer": "[{\"word\":\"gentlemen\",\"start\":0.17,\"end\":0.63}]"} {"utterance_id": "5639-40744-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0011.flac", "task_type": "lexical_detection", "query": "outside", "prompt": "When the utterance is complete, report all instances of the word \"outside\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.87, 2.43, "outside"]], "response_mode": "offline", "answer": "[{\"word\":\"outside\",\"start\":1.87,\"end\":2.43}]"} {"utterance_id": "2830-3980-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0039.flac", "task_type": "lexical_counting", "query": "terms", "prompt": "Listen to the entire clip, then report the count sequence for the word \"terms\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.63, 0.97, "one"], [2.27, 2.61, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.63,\"end\":0.97},{\"word\":\"two\",\"start\":2.27,\"end\":2.61}]"} {"utterance_id": "6829-68769-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0053.flac", "task_type": "lexical_detection", "query": "fifty", "prompt": "After the audio ends, extract all occurrences of the word \"fifty\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.24, 5.58, "fifty"]], "response_mode": "offline", "answer": "[{\"word\":\"fifty\",\"start\":5.24,\"end\":5.58}]"} {"utterance_id": "7021-79740-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0011.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[2.66, 3.21, "tiptoe"]], "response_mode": "offline", "answer": "[{\"word\":\"tiptoe\",\"start\":2.66,\"end\":3.21}]"} {"utterance_id": "237-126133-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0004.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.25, 2.46, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.25,\"end\":2.46}]"} {"utterance_id": "6930-76324-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0026.flac", "task_type": "lexical_detection", "query": "odd", "prompt": "After the audio ends, extract all occurrences of the word \"odd\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.03, 2.33, "odd"]], "response_mode": "offline", "answer": "[{\"word\":\"odd\",\"start\":2.03,\"end\":2.33}]"} {"utterance_id": "7021-85628-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0013.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.61, 7.99, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":7.61,\"end\":7.99}]"} {"utterance_id": "61-70968-0056", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0056.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.45, 2.93, "squire's"]], "response_mode": "offline", "answer": "[{\"word\":\"squire's\",\"start\":2.45,\"end\":2.93}]"} {"utterance_id": "2830-3980-0072", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0072.flac", "task_type": "lexical_detection", "query": "sold", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"sold\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.5, 3.84, "sold"]], "response_mode": "offline", "answer": "[{\"word\":\"sold\",\"start\":3.5,\"end\":3.84}]"} {"utterance_id": "6930-75918-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0019.flac", "task_type": "lexical_detection", "query": "tents", "prompt": "After the audio ends, extract all occurrences of the word \"tents\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.14, 4.59, "tents"]], "response_mode": "offline", "answer": "[{\"word\":\"tents\",\"start\":4.14,\"end\":4.59}]"} {"utterance_id": "61-70968-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0048.flac", "task_type": "lexical_detection", "query": "return", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"return\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.87, 1.21, "return"]], "response_mode": "offline", "answer": "[{\"word\":\"return\",\"start\":0.87,\"end\":1.21}]"} {"utterance_id": "1221-135767-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0004.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[8.84, 9.21, "eyes"], [12.71, 13.1, "hair"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":8.84,\"end\":9.21},{\"word\":\"hair\",\"start\":12.71,\"end\":13.1}]"} {"utterance_id": "5142-36377-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0022.flac", "task_type": "lexical_detection", "query": "wheeled", "prompt": "After listening to the whole recording, output every instance of the word \"wheeled\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.82, 8.11, "wheeled"]], "response_mode": "offline", "answer": "[{\"word\":\"wheeled\",\"start\":7.82,\"end\":8.11}]"} {"utterance_id": "5105-28241-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0004.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "You will hear a complete utterance. When it ends, list every instance of a weather term you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.68, 2.07, "breeze"]], "response_mode": "offline", "answer": "[{\"word\":\"breeze\",\"start\":1.68,\"end\":2.07}]"} {"utterance_id": "7021-79730-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0003.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.29, 2.85, "mary"]], "response_mode": "offline", "answer": "[{\"word\":\"mary\",\"start\":2.29,\"end\":2.85}]"} {"utterance_id": "8463-294828-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0022.flac", "task_type": "lexical_detection", "query": "abraham", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"abraham\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.06, 1.61, "abraham"]], "response_mode": "offline", "answer": "[{\"word\":\"abraham\",\"start\":1.06,\"end\":1.61}]"} {"utterance_id": "2830-3980-0073", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0073.flac", "task_type": "lexical_counting", "query": "conception", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"conception\" (word, start, end). If none, reply [].", "labels": [[2.13, 2.69, "one"], [3.87, 4.42, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.13,\"end\":2.69},{\"word\":\"two\",\"start\":3.87,\"end\":4.42}]"} {"utterance_id": "5142-33396-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0028.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.94, 2.33, "dozen"]], "response_mode": "offline", "answer": "[{\"word\":\"dozen\",\"start\":1.94,\"end\":2.33}]"} {"utterance_id": "237-134493-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0011.flac", "task_type": "lexical_detection", "query": "orchard", "prompt": "After the audio ends, extract all occurrences of the word \"orchard\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.39, 4.88, "orchard"]], "response_mode": "offline", "answer": "[{\"word\":\"orchard\",\"start\":4.39,\"end\":4.88}]"} {"utterance_id": "1284-1181-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0020.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.14, 4.69, "magician"]], "response_mode": "offline", "answer": "[{\"word\":\"magician\",\"start\":4.14,\"end\":4.69}]"} {"utterance_id": "1188-133604-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0003.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.67, 1.06, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":0.67,\"end\":1.06}]"} {"utterance_id": "2094-142345-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0033.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "Do not answer until the audio ends. Then list each instance of a disease name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.29, 1.97, "illness"]], "response_mode": "offline", "answer": "[{\"word\":\"illness\",\"start\":1.29,\"end\":1.97}]"} {"utterance_id": "4446-2273-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0029.flac", "task_type": "lexical_detection", "query": "room", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"room\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.32, 1.47, "room"]], "response_mode": "offline", "answer": "[{\"word\":\"room\",\"start\":1.32,\"end\":1.47}]"} {"utterance_id": "2961-961-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0022.flac", "task_type": "lexical_detection", "query": "explain", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"explain\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.23, 1.76, "explain"]], "response_mode": "offline", "answer": "[{\"word\":\"explain\",\"start\":1.23,\"end\":1.76}]"} {"utterance_id": "7176-88083-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0007.flac", "task_type": "lexical_detection", "query": "keenness", "prompt": "When the utterance is complete, report all instances of the word \"keenness\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.07, 1.55, "keenness"]], "response_mode": "offline", "answer": "[{\"word\":\"keenness\",\"start\":1.07,\"end\":1.55}]"} {"utterance_id": "8455-210777-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0059.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an organization name with fields word, start, end (seconds). If none, reply [].", "labels": [[4.74, 5.22, "government"]], "response_mode": "offline", "answer": "[{\"word\":\"government\",\"start\":4.74,\"end\":5.22}]"} {"utterance_id": "908-31957-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0015.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.22, 3.84, "crown"], [8.52, 9.0, "lips"]], "response_mode": "offline", "answer": "[{\"word\":\"crown\",\"start\":3.22,\"end\":3.84},{\"word\":\"lips\",\"start\":8.52,\"end\":9.0}]"} {"utterance_id": "3570-5695-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0007.flac", "task_type": "lexical_detection", "query": "yielded", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"yielded\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.14, 2.52, "yielded"]], "response_mode": "offline", "answer": "[{\"word\":\"yielded\",\"start\":2.14,\"end\":2.52}]"} {"utterance_id": "5142-33396-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0010.flac", "task_type": "lexical_detection", "query": "stern", "prompt": "After listening to the whole recording, output every instance of the word \"stern\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.37, 0.8, "stern"]], "response_mode": "offline", "answer": "[{\"word\":\"stern\",\"start\":0.37,\"end\":0.8}]"} {"utterance_id": "7729-102255-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0022.flac", "task_type": "lexical_detection", "query": "sprang", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"sprang\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.34, 1.83, "sprang"]], "response_mode": "offline", "answer": "[{\"word\":\"sprang\",\"start\":1.34,\"end\":1.83}]"} {"utterance_id": "7729-102255-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0035.flac", "task_type": "lexical_detection", "query": "meanwhile", "prompt": "Process the full audio offline. Find every instance of the word \"meanwhile\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.87, 4.35, "meanwhile"]], "response_mode": "offline", "answer": "[{\"word\":\"meanwhile\",\"start\":3.87,\"end\":4.35}]"} {"utterance_id": "1580-141083-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0003.flac", "task_type": "lexical_detection", "query": "homely", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"homely\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.89, 3.37, "homely"]], "response_mode": "offline", "answer": "[{\"word\":\"homely\",\"start\":2.89,\"end\":3.37}]"} {"utterance_id": "4992-23283-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0009.flac", "task_type": "lexical_detection", "query": "woodley", "prompt": "Process the full audio offline. Find every instance of the word \"woodley\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.02, 2.36, "woodley"]], "response_mode": "offline", "answer": "[{\"word\":\"woodley\",\"start\":2.02,\"end\":2.36}]"} {"utterance_id": "672-122797-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0026.flac", "task_type": "semantic_detection", "query": "money", "prompt": "After the audio ends, extract all occurrences of a money or currency term. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.22, 10.88, "crowns"]], "response_mode": "offline", "answer": "[{\"word\":\"crowns\",\"start\":10.22,\"end\":10.88}]"} {"utterance_id": "5683-32866-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0011.flac", "task_type": "lexical_detection", "query": "satisfied", "prompt": "When the utterance is complete, report all instances of the word \"satisfied\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.94, 4.57, "satisfied"]], "response_mode": "offline", "answer": "[{\"word\":\"satisfied\",\"start\":3.94,\"end\":4.57}]"} {"utterance_id": "8230-279154-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0037.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "After the audio ends, extract all occurrences of an animal name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.53, 12.09, "cat"], [18.81, 19.28, "cat"]], "response_mode": "offline", "answer": "[{\"word\":\"cat\",\"start\":11.53,\"end\":12.09},{\"word\":\"cat\",\"start\":18.81,\"end\":19.28}]"} {"utterance_id": "8463-294828-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0025.flac", "task_type": "lexical_detection", "query": "we're", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"we're\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.37, 0.5, "we're"]], "response_mode": "offline", "answer": "[{\"word\":\"we're\",\"start\":0.37,\"end\":0.5}]"} {"utterance_id": "7021-79740-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0008.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Once the audio has finished, identify every instance of a family relation. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.98, 3.46, "children"]], "response_mode": "offline", "answer": "[{\"word\":\"children\",\"start\":2.98,\"end\":3.46}]"} {"utterance_id": "4446-2271-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0006.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.65, 1.89, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":1.65,\"end\":1.89}]"} {"utterance_id": "1284-1181-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0011.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.29, 1.86, "magician"]], "response_mode": "offline", "answer": "[{\"word\":\"magician\",\"start\":1.29,\"end\":1.86}]"} {"utterance_id": "4446-2273-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0021.flac", "task_type": "semantic_detection", "query": "money", "prompt": "When the utterance is complete, report all instances of a money or currency term including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.99, 3.54, "francs"]], "response_mode": "offline", "answer": "[{\"word\":\"francs\",\"start\":2.99,\"end\":3.54}]"} {"utterance_id": "908-31957-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0007.flac", "task_type": "lexical_detection", "query": "set", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"set\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.75, 3.25, "set"]], "response_mode": "offline", "answer": "[{\"word\":\"set\",\"start\":2.75,\"end\":3.25}]"} {"utterance_id": "1089-134686-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0001.flac", "task_type": "lexical_detection", "query": "counselled", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"counselled\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.12, 2.64, "counselled"]], "response_mode": "offline", "answer": "[{\"word\":\"counselled\",\"start\":2.12,\"end\":2.64}]"} {"utterance_id": "2830-3980-0061", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0061.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "After the audio ends, extract all occurrences of an animal name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.51, 3.04, "lambs"]], "response_mode": "offline", "answer": "[{\"word\":\"lambs\",\"start\":2.51,\"end\":3.04}]"} {"utterance_id": "6930-75918-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0014.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.96, 2.45, "raoul"]], "response_mode": "offline", "answer": "[{\"word\":\"raoul\",\"start\":1.96,\"end\":2.45}]"} {"utterance_id": "237-126133-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0014.flac", "task_type": "lexical_detection", "query": "jasper's", "prompt": "When the utterance is complete, report all instances of the word \"jasper's\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.61, 6.09, "jasper's"]], "response_mode": "offline", "answer": "[{\"word\":\"jasper's\",\"start\":5.61,\"end\":6.09}]"} {"utterance_id": "260-123286-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0029.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen carefully to the entire clip, then report every instance of an animal name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.04, 1.51, "creatures"]], "response_mode": "offline", "answer": "[{\"word\":\"creatures\",\"start\":1.04,\"end\":1.51}]"} {"utterance_id": "2830-3979-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0004.flac", "task_type": "lexical_detection", "query": "latin", "prompt": "After listening to the whole recording, output every instance of the word \"latin\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.18, 1.69, "latin"]], "response_mode": "offline", "answer": "[{\"word\":\"latin\",\"start\":1.18,\"end\":1.69}]"} {"utterance_id": "237-134500-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0006.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "When the utterance is complete, report all instances of a weather term including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.66, 4.04, "rain"]], "response_mode": "offline", "answer": "[{\"word\":\"rain\",\"start\":3.66,\"end\":4.04}]"} {"utterance_id": "4077-13751-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0010.flac", "task_type": "lexical_detection", "query": "building", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"building\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.02, 4.4, "building"]], "response_mode": "offline", "answer": "[{\"word\":\"building\",\"start\":4.02,\"end\":4.4}]"} {"utterance_id": "8455-210777-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0017.flac", "task_type": "lexical_detection", "query": "dinner", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"dinner\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.47, 4.85, "dinner"]], "response_mode": "offline", "answer": "[{\"word\":\"dinner\",\"start\":4.47,\"end\":4.85}]"} {"utterance_id": "7176-92135-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0043.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "After listening to the whole recording, output timed count words for a number (one, two, three, ...) as JSON. If none, reply [].", "labels": [[0.25, 0.4, "one"], [8.49, 8.75, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.25,\"end\":0.4},{\"word\":\"two\",\"start\":8.49,\"end\":8.75}]"} {"utterance_id": "4446-2275-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0030.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.59, 0.87, "hilda"]], "response_mode": "offline", "answer": "[{\"word\":\"hilda\",\"start\":0.59,\"end\":0.87}]"} {"utterance_id": "8230-279154-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0038.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen to the full audio. After it finishes, find every instance of an animal name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[15.51, 16.1, "animals"]], "response_mode": "offline", "answer": "[{\"word\":\"animals\",\"start\":15.51,\"end\":16.1}]"} {"utterance_id": "8555-284447-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0014.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen to the full audio. After it finishes, find every instance of an animal name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.33, 2.7, "goat"]], "response_mode": "offline", "answer": "[{\"word\":\"goat\",\"start\":2.33,\"end\":2.7}]"} {"utterance_id": "7021-85628-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0005.flac", "task_type": "lexical_detection", "query": "visit", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"visit\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.87, 4.13, "visit"]], "response_mode": "offline", "answer": "[{\"word\":\"visit\",\"start\":3.87,\"end\":4.13}]"} {"utterance_id": "1188-133604-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0008.flac", "task_type": "lexical_detection", "query": "philosophy", "prompt": "After listening to the whole recording, output every instance of the word \"philosophy\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.46, 3.06, "philosophy"]], "response_mode": "offline", "answer": "[{\"word\":\"philosophy\",\"start\":2.46,\"end\":3.06}]"} {"utterance_id": "8463-294828-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0005.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.4, 2.38, "manservant"]], "response_mode": "offline", "answer": "[{\"word\":\"manservant\",\"start\":1.4,\"end\":2.38}]"} {"utterance_id": "237-126133-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0020.flac", "task_type": "lexical_detection", "query": "let", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"let\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.35, 1.52, "let"]], "response_mode": "offline", "answer": "[{\"word\":\"let\",\"start\":1.35,\"end\":1.52}]"} {"utterance_id": "3729-6852-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0014.flac", "task_type": "lexical_detection", "query": "louis", "prompt": "After the audio ends, extract all occurrences of the word \"louis\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.24, 3.81, "louis"]], "response_mode": "offline", "answer": "[{\"word\":\"louis\",\"start\":3.24,\"end\":3.81}]"} {"utterance_id": "7729-102255-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0000.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After the audio ends, extract all occurrences of an organization name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.69, 1.45, "legislature"]], "response_mode": "offline", "answer": "[{\"word\":\"legislature\",\"start\":0.69,\"end\":1.45}]"} {"utterance_id": "5142-36377-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0004.flac", "task_type": "lexical_detection", "query": "left", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"left\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.42, 4.72, "left"]], "response_mode": "offline", "answer": "[{\"word\":\"left\",\"start\":4.42,\"end\":4.72}]"} {"utterance_id": "5683-32879-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0005.flac", "task_type": "lexical_detection", "query": "transient", "prompt": "Once the audio has finished, identify every instance of the word \"transient\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.54, 1.24, "transient"]], "response_mode": "offline", "answer": "[{\"word\":\"transient\",\"start\":0.54,\"end\":1.24}]"} {"utterance_id": "6829-68771-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0001.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.22, 0.69, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.22,\"end\":0.69}]"} {"utterance_id": "1284-1180-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0000.flac", "task_type": "lexical_counting", "query": "gold", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"gold\" with word/start/end. If none, reply [].", "labels": [[3.13, 3.44, "one"], [7.48, 7.94, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.13,\"end\":3.44},{\"word\":\"two\",\"start\":7.48,\"end\":7.94}]"} {"utterance_id": "1188-133604-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0023.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.27, 2.61, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":2.27,\"end\":2.61}]"} {"utterance_id": "237-134493-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0001.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the full audio. After it finishes, find every instance of a family relation. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.57, 1.09, "wife"]], "response_mode": "offline", "answer": "[{\"word\":\"wife\",\"start\":0.57,\"end\":1.09}]"} {"utterance_id": "2094-142345-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0028.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.0, 2.36, "totty"], [3.9, 4.86, "missus poyser"]], "response_mode": "offline", "answer": "[{\"word\":\"totty\",\"start\":2.0,\"end\":2.36},{\"word\":\"missus poyser\",\"start\":3.9,\"end\":4.86}]"} {"utterance_id": "3729-6852-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0037.flac", "task_type": "lexical_detection", "query": "introduced", "prompt": "Once the audio has finished, identify every instance of the word \"introduced\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.58, 1.09, "introduced"]], "response_mode": "offline", "answer": "[{\"word\":\"introduced\",\"start\":0.58,\"end\":1.09}]"} {"utterance_id": "6829-68769-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0045.flac", "task_type": "lexical_detection", "query": "pay", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"pay\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.43, 0.73, "pay"]], "response_mode": "offline", "answer": "[{\"word\":\"pay\",\"start\":0.43,\"end\":0.73}]"} {"utterance_id": "8455-210777-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0024.flac", "task_type": "lexical_detection", "query": "during", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"during\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.41, 5.68, "during"]], "response_mode": "offline", "answer": "[{\"word\":\"during\",\"start\":5.41,\"end\":5.68}]"} {"utterance_id": "2300-131720-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0033.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[9.93, 10.29, "six"]], "response_mode": "offline", "answer": "[{\"word\":\"six\",\"start\":9.93,\"end\":10.29}]"} {"utterance_id": "8224-274381-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0014.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.6, 5.23, "officers"], [7.23, 7.95, "generals"]], "response_mode": "offline", "answer": "[{\"word\":\"officers\",\"start\":4.6,\"end\":5.23},{\"word\":\"generals\",\"start\":7.23,\"end\":7.95}]"} {"utterance_id": "5683-32865-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0011.flac", "task_type": "lexical_detection", "query": "whereupon", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"whereupon\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.21, 0.95, "whereupon"]], "response_mode": "offline", "answer": "[{\"word\":\"whereupon\",\"start\":0.21,\"end\":0.95}]"} {"utterance_id": "1995-1826-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0024.flac", "task_type": "lexical_detection", "query": "fleece", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"fleece\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.87, 1.57, "fleece"], [2.85, 3.51, "fleece"]], "response_mode": "offline", "answer": "[{\"word\":\"fleece\",\"start\":0.87,\"end\":1.57},{\"word\":\"fleece\",\"start\":2.85,\"end\":3.51}]"} {"utterance_id": "4507-16021-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0036.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.63, 2.08, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.63,\"end\":2.08}]"} {"utterance_id": "61-70970-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0021.flac", "task_type": "lexical_detection", "query": "journey", "prompt": "Process the full audio offline. Find every instance of the word \"journey\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.22, 1.73, "journey"]], "response_mode": "offline", "answer": "[{\"word\":\"journey\",\"start\":1.22,\"end\":1.73}]"} {"utterance_id": "61-70968-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0000.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.01, 2.41, "wizard"]], "response_mode": "offline", "answer": "[{\"word\":\"wizard\",\"start\":2.01,\"end\":2.41}]"} {"utterance_id": "5142-36377-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0003.flac", "task_type": "lexical_detection", "query": "ambrose", "prompt": "After listening to the whole recording, output every instance of the word \"ambrose\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.39, 0.89, "ambrose"]], "response_mode": "offline", "answer": "[{\"word\":\"ambrose\",\"start\":0.39,\"end\":0.89}]"} {"utterance_id": "908-157963-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0005.flac", "task_type": "lexical_detection", "query": "music", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"music\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.23, 4.76, "music"]], "response_mode": "offline", "answer": "[{\"word\":\"music\",\"start\":4.23,\"end\":4.76}]"} {"utterance_id": "3570-5694-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0010.flac", "task_type": "lexical_counting", "query": "exception", "prompt": "When the utterance is complete, report how the word \"exception\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.38, 6.85, "one"], [9.91, 10.41, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.38,\"end\":6.85},{\"word\":\"two\",\"start\":9.91,\"end\":10.41}]"} {"utterance_id": "4970-29093-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0001.flac", "task_type": "lexical_detection", "query": "horizon", "prompt": "When the utterance is complete, report all instances of the word \"horizon\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.78, 11.49, "horizon"]], "response_mode": "offline", "answer": "[{\"word\":\"horizon\",\"start\":10.78,\"end\":11.49}]"} {"utterance_id": "1188-133604-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0030.flac", "task_type": "lexical_detection", "query": "knows", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"knows\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.64, 0.88, "knows"]], "response_mode": "offline", "answer": "[{\"word\":\"knows\",\"start\":0.64,\"end\":0.88}]"} {"utterance_id": "8463-287645-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0012.flac", "task_type": "lexical_detection", "query": "jacob's", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"jacob's\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.45, 1.94, "jacob's"]], "response_mode": "offline", "answer": "[{\"word\":\"jacob's\",\"start\":1.45,\"end\":1.94}]"} {"utterance_id": "6829-68769-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0012.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.44, 2.94, "markham"]], "response_mode": "offline", "answer": "[{\"word\":\"markham\",\"start\":2.44,\"end\":2.94}]"} {"utterance_id": "4970-29093-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0009.flac", "task_type": "lexical_detection", "query": "planned", "prompt": "After the audio ends, extract all occurrences of the word \"planned\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.99, 4.36, "planned"]], "response_mode": "offline", "answer": "[{\"word\":\"planned\",\"start\":3.99,\"end\":4.36}]"} {"utterance_id": "672-122797-0061", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0061.flac", "task_type": "semantic_detection", "query": "food", "prompt": "After the audio ends, extract all occurrences of a food name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.88, 2.34, "bacon"]], "response_mode": "offline", "answer": "[{\"word\":\"bacon\",\"start\":1.88,\"end\":2.34}]"} {"utterance_id": "672-122797-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0004.flac", "task_type": "lexical_detection", "query": "leaping", "prompt": "After listening to the whole recording, output every instance of the word \"leaping\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.92, 6.36, "leaping"]], "response_mode": "offline", "answer": "[{\"word\":\"leaping\",\"start\":5.92,\"end\":6.36}]"} {"utterance_id": "7176-92135-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0026.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Do not answer until the audio ends. Then list each instance of an animal name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.25, 2.8, "boar hound"]], "response_mode": "offline", "answer": "[{\"word\":\"boar hound\",\"start\":2.25,\"end\":2.8}]"} {"utterance_id": "3575-170457-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0051.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.21, 4.62, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":4.21,\"end\":4.62}]"} {"utterance_id": "1221-135767-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0023.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an ordinal number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.41, 8.73, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":8.41,\"end\":8.73}]"} {"utterance_id": "1221-135766-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0014.flac", "task_type": "lexical_detection", "query": "gazed", "prompt": "Once the audio has finished, identify every instance of the word \"gazed\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.61, 2.0, "gazed"]], "response_mode": "offline", "answer": "[{\"word\":\"gazed\",\"start\":1.61,\"end\":2.0}]"} {"utterance_id": "8555-284447-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0001.flac", "task_type": "lexical_detection", "query": "fast", "prompt": "After listening to the whole recording, output every instance of the word \"fast\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.79, 1.18, "fast"]], "response_mode": "offline", "answer": "[{\"word\":\"fast\",\"start\":0.79,\"end\":1.18}]"} {"utterance_id": "7729-102255-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0018.flac", "task_type": "lexical_detection", "query": "territorial", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"territorial\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.86, 8.7, "territorial"]], "response_mode": "offline", "answer": "[{\"word\":\"territorial\",\"start\":7.86,\"end\":8.7}]"} {"utterance_id": "8463-294828-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0033.flac", "task_type": "lexical_detection", "query": "cabin", "prompt": "After the audio ends, extract all occurrences of the word \"cabin\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.78, 2.25, "cabin"]], "response_mode": "offline", "answer": "[{\"word\":\"cabin\",\"start\":1.78,\"end\":2.25}]"} {"utterance_id": "5142-36377-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0021.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Once the audio has finished, identify every instance of a food name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[18.27, 18.65, "supper"]], "response_mode": "offline", "answer": "[{\"word\":\"supper\",\"start\":18.27,\"end\":18.65}]"} {"utterance_id": "6829-68769-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.78, 3.28, "markham"]], "response_mode": "offline", "answer": "[{\"word\":\"markham\",\"start\":2.78,\"end\":3.28}]"} {"utterance_id": "1580-141083-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0046.flac", "task_type": "lexical_detection", "query": "times", "prompt": "After the audio ends, extract all occurrences of the word \"times\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.43, 3.04, "times"]], "response_mode": "offline", "answer": "[{\"word\":\"times\",\"start\":2.43,\"end\":3.04}]"} {"utterance_id": "8455-210777-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0014.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.54, 1.67, "sir kennington oval"]], "response_mode": "offline", "answer": "[{\"word\":\"sir kennington oval\",\"start\":0.54,\"end\":1.67}]"} {"utterance_id": "6930-76324-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0025.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.92, 1.63, "goliath"]], "response_mode": "offline", "answer": "[{\"word\":\"goliath\",\"start\":0.92,\"end\":1.63}]"} {"utterance_id": "1188-133604-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0022.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[1.44, 2.18, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":1.44,\"end\":2.18}]"} {"utterance_id": "2830-3979-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0002.flac", "task_type": "lexical_detection", "query": "commentary", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"commentary\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.95, "commentary"]], "response_mode": "offline", "answer": "[{\"word\":\"commentary\",\"start\":2.35,\"end\":2.95}]"} {"utterance_id": "3575-170457-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0029.flac", "task_type": "lexical_detection", "query": "pleasure", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"pleasure\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.95, 2.47, "pleasure"]], "response_mode": "offline", "answer": "[{\"word\":\"pleasure\",\"start\":1.95,\"end\":2.47}]"} {"utterance_id": "4446-2273-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0016.flac", "task_type": "lexical_detection", "query": "nose", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"nose\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[6.77, 7.18, "nose"]], "response_mode": "offline", "answer": "[{\"word\":\"nose\",\"start\":6.77,\"end\":7.18}]"} {"utterance_id": "1089-134686-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0022.flac", "task_type": "lexical_detection", "query": "beatitude", "prompt": "After the audio ends, extract all occurrences of the word \"beatitude\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.21, 2.91, "beatitude"], [6.5, 7.22, "beatitude"]], "response_mode": "offline", "answer": "[{\"word\":\"beatitude\",\"start\":2.21,\"end\":2.91},{\"word\":\"beatitude\",\"start\":6.5,\"end\":7.22}]"} {"utterance_id": "4970-29095-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0031.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Process the full audio offline. Find every instance of a money or currency term and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.84, 11.29, "fortunes"]], "response_mode": "offline", "answer": "[{\"word\":\"fortunes\",\"start\":10.84,\"end\":11.29}]"} {"utterance_id": "3570-5694-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0016.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.94, 13.21, "servants retainers"]], "response_mode": "offline", "answer": "[{\"word\":\"servants retainers\",\"start\":11.94,\"end\":13.21}]"} {"utterance_id": "8230-279154-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0025.flac", "task_type": "lexical_detection", "query": "expectations", "prompt": "When the utterance is complete, report all instances of the word \"expectations\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[16.34, 17.3, "expectations"]], "response_mode": "offline", "answer": "[{\"word\":\"expectations\",\"start\":16.34,\"end\":17.3}]"} {"utterance_id": "8555-284447-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0009.flac", "task_type": "lexical_detection", "query": "mornin", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"mornin\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.24, 0.73, "mornin"]], "response_mode": "offline", "answer": "[{\"word\":\"mornin\",\"start\":0.24,\"end\":0.73}]"} {"utterance_id": "121-127105-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0024.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 1.0, "douglas"]], "response_mode": "offline", "answer": "[{\"word\":\"douglas\",\"start\":0.45,\"end\":1.0}]"} {"utterance_id": "3570-5694-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0014.flac", "task_type": "lexical_detection", "query": "affiliated", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"affiliated\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.9, 1.53, "affiliated"]], "response_mode": "offline", "answer": "[{\"word\":\"affiliated\",\"start\":0.9,\"end\":1.53}]"} {"utterance_id": "3729-6852-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0015.flac", "task_type": "lexical_detection", "query": "service", "prompt": "Process the full audio offline. Find every instance of the word \"service\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.02, 1.52, "service"]], "response_mode": "offline", "answer": "[{\"word\":\"service\",\"start\":1.02,\"end\":1.52}]"} {"utterance_id": "5142-33396-0049", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0049.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.76, "farmer's"]], "response_mode": "offline", "answer": "[{\"word\":\"farmer's\",\"start\":2.35,\"end\":2.76}]"} {"utterance_id": "2830-3980-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0052.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.01, 4.42, "teacher"]], "response_mode": "offline", "answer": "[{\"word\":\"teacher\",\"start\":4.01,\"end\":4.42}]"} {"utterance_id": "1188-133604-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0013.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.57, 1.9, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.57,\"end\":1.9}]"} {"utterance_id": "6930-76324-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0005.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "You will hear a complete utterance. When it ends, list every instance of a family relation you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.84, 1.14, "brother"]], "response_mode": "offline", "answer": "[{\"word\":\"brother\",\"start\":0.84,\"end\":1.14}]"} {"utterance_id": "260-123286-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0004.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.47, 0.69, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.47,\"end\":0.69}]"} {"utterance_id": "121-121726-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0014.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.44, 2.93, "dealer"]], "response_mode": "offline", "answer": "[{\"word\":\"dealer\",\"start\":2.44,\"end\":2.93}]"} {"utterance_id": "8555-284447-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0000.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Listen carefully to the entire clip, then report every instance of a state or province name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.69, 7.25, "dominions"]], "response_mode": "offline", "answer": "[{\"word\":\"dominions\",\"start\":6.69,\"end\":7.25}]"} {"utterance_id": "8463-294828-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0019.flac", "task_type": "semantic_detection", "query": "country", "prompt": "You will hear a complete utterance. When it ends, list every instance of a country name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.66, 4.25, "france"]], "response_mode": "offline", "answer": "[{\"word\":\"france\",\"start\":3.66,\"end\":4.25}]"} {"utterance_id": "237-134500-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0013.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.05, 4.43, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.05,\"end\":4.43}]"} {"utterance_id": "61-70968-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0045.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.26, 2.66, "sheriff's"]], "response_mode": "offline", "answer": "[{\"word\":\"sheriff's\",\"start\":2.26,\"end\":2.66}]"} {"utterance_id": "908-157963-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0020.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.69, 9.35, "flowers"]], "response_mode": "offline", "answer": "[{\"word\":\"flowers\",\"start\":8.69,\"end\":9.35}]"} {"utterance_id": "4077-13751-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0001.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.07, 4.64, "seed"]], "response_mode": "offline", "answer": "[{\"word\":\"seed\",\"start\":4.07,\"end\":4.64}]"} {"utterance_id": "1284-1180-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.4, 0.72, "unc"]], "response_mode": "offline", "answer": "[{\"word\":\"unc\",\"start\":0.4,\"end\":0.72}]"} {"utterance_id": "6930-75918-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0002.flac", "task_type": "lexical_detection", "query": "poured", "prompt": "When the utterance is complete, report all instances of the word \"poured\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.49, 1.87, "poured"]], "response_mode": "offline", "answer": "[{\"word\":\"poured\",\"start\":1.49,\"end\":1.87}]"} {"utterance_id": "1320-122612-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0002.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.45, 1.64, "few"]], "response_mode": "offline", "answer": "[{\"word\":\"few\",\"start\":1.45,\"end\":1.64}]"} {"utterance_id": "6930-76324-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0017.flac", "task_type": "lexical_detection", "query": "urged", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"urged\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.28, 2.79, "urged"]], "response_mode": "offline", "answer": "[{\"word\":\"urged\",\"start\":2.28,\"end\":2.79}]"} {"utterance_id": "5105-28241-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0013.flac", "task_type": "semantic_detection", "query": "city", "prompt": "When the utterance is complete, report all instances of a city name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.08, 3.46, "city"]], "response_mode": "offline", "answer": "[{\"word\":\"city\",\"start\":3.08,\"end\":3.46}]"} {"utterance_id": "8463-294828-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0028.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Do not answer until the audio ends. Then list each instance of an animal name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.59, 3.41, "babirusa"]], "response_mode": "offline", "answer": "[{\"word\":\"babirusa\",\"start\":2.59,\"end\":3.41}]"} {"utterance_id": "3575-170457-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0045.flac", "task_type": "lexical_detection", "query": "intimate", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"intimate\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.01, 5.43, "intimate"]], "response_mode": "offline", "answer": "[{\"word\":\"intimate\",\"start\":5.01,\"end\":5.43}]"} {"utterance_id": "7176-88083-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0012.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "When the utterance is complete, report all instances of a plant name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.59, 2.0, "branch"]], "response_mode": "offline", "answer": "[{\"word\":\"branch\",\"start\":1.59,\"end\":2.0}]"} {"utterance_id": "2961-960-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0005.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.27, 5.03, "philosopher"]], "response_mode": "offline", "answer": "[{\"word\":\"philosopher\",\"start\":4.27,\"end\":5.03}]"} {"utterance_id": "7127-75947-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0026.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.22, 2.76, "mademoiselle de tonnay charente"]], "response_mode": "offline", "answer": "[{\"word\":\"mademoiselle de tonnay charente\",\"start\":1.22,\"end\":2.76}]"} {"utterance_id": "1995-1836-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0005.flac", "task_type": "lexical_detection", "query": "grey", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"grey\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 0.83, "grey"]], "response_mode": "offline", "answer": "[{\"word\":\"grey\",\"start\":0.52,\"end\":0.83}]"} {"utterance_id": "5142-33396-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0004.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.4, 1.99, "brothers"]], "response_mode": "offline", "answer": "[{\"word\":\"brothers\",\"start\":1.4,\"end\":1.99}]"} {"utterance_id": "8224-274381-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0016.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.16, 4.49, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.16,\"end\":4.49}]"} {"utterance_id": "4077-13751-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0003.flac", "task_type": "lexical_detection", "query": "smallest", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"smallest\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.53, 2.99, "smallest"]], "response_mode": "offline", "answer": "[{\"word\":\"smallest\",\"start\":2.53,\"end\":2.99}]"} {"utterance_id": "4077-13754-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0011.flac", "task_type": "lexical_detection", "query": "utah", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"utah\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.68, 3.12, "utah"]], "response_mode": "offline", "answer": "[{\"word\":\"utah\",\"start\":2.68,\"end\":3.12}]"} {"utterance_id": "5105-28240-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0019.flac", "task_type": "lexical_detection", "query": "yacht", "prompt": "Process the full audio offline. Find every instance of the word \"yacht\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.84, 1.21, "yacht"]], "response_mode": "offline", "answer": "[{\"word\":\"yacht\",\"start\":0.84,\"end\":1.21}]"} {"utterance_id": "3570-5696-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0005.flac", "task_type": "semantic_counting", "query": "money", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a money or currency term: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[2.38, 3.13, "one"], [11.13, 11.79, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.38,\"end\":3.13},{\"word\":\"two\",\"start\":11.13,\"end\":11.79}]"} {"utterance_id": "5105-28241-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0001.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the full audio. After it finishes, find every instance of an organization name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.73, 4.12, "navy"]], "response_mode": "offline", "answer": "[{\"word\":\"navy\",\"start\":3.73,\"end\":4.12}]"} {"utterance_id": "5639-40744-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0021.flac", "task_type": "lexical_detection", "query": "lest", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"lest\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[8.09, 8.4, "lest"]], "response_mode": "offline", "answer": "[{\"word\":\"lest\",\"start\":8.09,\"end\":8.4}]"} {"utterance_id": "908-31957-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0022.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Do not answer until the audio ends. Then list each instance of a weather term as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[11.93, 12.59, "dews"]], "response_mode": "offline", "answer": "[{\"word\":\"dews\",\"start\":11.93,\"end\":12.59}]"} {"utterance_id": "61-70970-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0031.flac", "task_type": "lexical_detection", "query": "waving", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"waving\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.93, 1.32, "waving"]], "response_mode": "offline", "answer": "[{\"word\":\"waving\",\"start\":0.93,\"end\":1.32}]"} {"utterance_id": "1221-135767-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.97, 9.43, "hester"]], "response_mode": "offline", "answer": "[{\"word\":\"hester\",\"start\":8.97,\"end\":9.43}]"} {"utterance_id": "1284-1180-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0028.flac", "task_type": "lexical_detection", "query": "quilt", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"quilt\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.88, 1.34, "quilt"]], "response_mode": "offline", "answer": "[{\"word\":\"quilt\",\"start\":0.88,\"end\":1.34}]"} {"utterance_id": "4507-16021-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0023.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "After the audio ends, tally every instance of a profession. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.91, 2.05, "one"], [9.5, 10.61, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.91,\"end\":2.05},{\"word\":\"two\",\"start\":9.5,\"end\":10.61}]"} {"utterance_id": "121-127105-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0035.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.38, 8.92, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":8.38,\"end\":8.92}]"} {"utterance_id": "2094-142345-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0003.flac", "task_type": "lexical_counting", "query": "floor", "prompt": "Listen to the entire clip, then report the count sequence for the word \"floor\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.02, 5.62, "one"], [11.93, 12.43, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.02,\"end\":5.62},{\"word\":\"two\",\"start\":11.93,\"end\":12.43}]"} {"utterance_id": "7176-88083-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0024.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Process the full audio offline. Find every instance of an animal name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.31, 1.64, "fly"], [4.63, 5.06, "hawk's"]], "response_mode": "offline", "answer": "[{\"word\":\"fly\",\"start\":1.31,\"end\":1.64},{\"word\":\"hawk's\",\"start\":4.63,\"end\":5.06}]"} {"utterance_id": "6829-68769-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0005.flac", "task_type": "lexical_detection", "query": "impetuous", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"impetuous\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.72, 2.42, "impetuous"]], "response_mode": "offline", "answer": "[{\"word\":\"impetuous\",\"start\":1.72,\"end\":2.42}]"} {"utterance_id": "7127-75947-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0033.flac", "task_type": "lexical_detection", "query": "spoke", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"spoke\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.08, 6.52, "spoke"]], "response_mode": "offline", "answer": "[{\"word\":\"spoke\",\"start\":6.08,\"end\":6.52}]"} {"utterance_id": "5142-33396-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0054.flac", "task_type": "lexical_detection", "query": "point", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"point\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.79, 3.13, "point"]], "response_mode": "offline", "answer": "[{\"word\":\"point\",\"start\":2.79,\"end\":3.13}]"} {"utterance_id": "8463-294828-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0034.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.54, 2.99, "conseil"]], "response_mode": "offline", "answer": "[{\"word\":\"conseil\",\"start\":2.54,\"end\":2.99}]"} {"utterance_id": "4970-29093-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0006.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.08, 14.39, "philip"]], "response_mode": "offline", "answer": "[{\"word\":\"philip\",\"start\":14.08,\"end\":14.39}]"} {"utterance_id": "4992-41797-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0019.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a person's name: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.54, 1.16, "one"], [2.62, 3.08, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.54,\"end\":1.16},{\"word\":\"two\",\"start\":2.62,\"end\":3.08}]"} {"utterance_id": "3575-170457-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0051.flac", "task_type": "lexical_detection", "query": "tea", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"tea\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.45, 0.62, "tea"]], "response_mode": "offline", "answer": "[{\"word\":\"tea\",\"start\":0.45,\"end\":0.62}]"} {"utterance_id": "1089-134691-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0006.flac", "task_type": "lexical_detection", "query": "dim", "prompt": "After the audio ends, extract all occurrences of the word \"dim\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.95, 1.22, "dim"]], "response_mode": "offline", "answer": "[{\"word\":\"dim\",\"start\":0.95,\"end\":1.22}]"} {"utterance_id": "5142-36377-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0015.flac", "task_type": "lexical_counting", "query": "impressions", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"impressions\" (word, start, end). If none, reply [].", "labels": [[0.54, 0.98, "one"], [3.52, 4.21, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.54,\"end\":0.98},{\"word\":\"two\",\"start\":3.52,\"end\":4.21}]"} {"utterance_id": "61-70968-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0021.flac", "task_type": "lexical_detection", "query": "submit", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"submit\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.98, 1.33, "submit"]], "response_mode": "offline", "answer": "[{\"word\":\"submit\",\"start\":0.98,\"end\":1.33}]"} {"utterance_id": "8455-210777-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0012.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.47, 1.39, "missus neverbend"]], "response_mode": "offline", "answer": "[{\"word\":\"missus neverbend\",\"start\":0.47,\"end\":1.39}]"} {"utterance_id": "260-123288-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0013.flac", "task_type": "lexical_detection", "query": "vapours", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"vapours\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.83, 1.31, "vapours"]], "response_mode": "offline", "answer": "[{\"word\":\"vapours\",\"start\":0.83,\"end\":1.31}]"} {"utterance_id": "5142-33396-0065", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0065.flac", "task_type": "lexical_detection", "query": "gently", "prompt": "Once the audio has finished, identify every instance of the word \"gently\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.46, 1.79, "gently"]], "response_mode": "offline", "answer": "[{\"word\":\"gently\",\"start\":1.46,\"end\":1.79}]"} {"utterance_id": "2094-142345-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0030.flac", "task_type": "lexical_detection", "query": "descended", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"descended\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.85, 2.3, "descended"]], "response_mode": "offline", "answer": "[{\"word\":\"descended\",\"start\":1.85,\"end\":2.3}]"} {"utterance_id": "8230-279154-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0015.flac", "task_type": "lexical_detection", "query": "remote", "prompt": "Process the full audio offline. Find every instance of the word \"remote\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.62, 6.25, "remote"]], "response_mode": "offline", "answer": "[{\"word\":\"remote\",\"start\":5.62,\"end\":6.25}]"} {"utterance_id": "7729-102255-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0012.flac", "task_type": "lexical_detection", "query": "summons", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"summons\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.3, 3.91, "summons"]], "response_mode": "offline", "answer": "[{\"word\":\"summons\",\"start\":3.3,\"end\":3.91}]"} {"utterance_id": "4992-41797-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0018.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "After listening to the whole recording, output every instance of a disease name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.32, 8.97, "injury"]], "response_mode": "offline", "answer": "[{\"word\":\"injury\",\"start\":8.32,\"end\":8.97}]"} {"utterance_id": "5683-32866-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0013.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.89, 9.5, "features"]], "response_mode": "offline", "answer": "[{\"word\":\"features\",\"start\":8.89,\"end\":9.5}]"} {"utterance_id": "8230-279154-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0042.flac", "task_type": "lexical_detection", "query": "virtue", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"virtue\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.34, 5.88, "virtue"]], "response_mode": "offline", "answer": "[{\"word\":\"virtue\",\"start\":5.34,\"end\":5.88}]"} {"utterance_id": "7021-85628-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0024.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.67, 1.05, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":0.67,\"end\":1.05}]"} {"utterance_id": "2830-3980-0058", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0058.flac", "task_type": "lexical_detection", "query": "christ", "prompt": "Once the audio has finished, identify every instance of the word \"christ\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.96, 2.42, "christ"]], "response_mode": "offline", "answer": "[{\"word\":\"christ\",\"start\":1.96,\"end\":2.42}]"} {"utterance_id": "2094-142345-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0039.flac", "task_type": "lexical_counting", "query": "i've", "prompt": "Listen to the entire clip, then report the count sequence for the word \"i've\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.26, 0.45, "one"], [5.39, 5.58, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.26,\"end\":0.45},{\"word\":\"two\",\"start\":5.39,\"end\":5.58}]"} {"utterance_id": "6829-68771-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0029.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.75, 2.95, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":2.75,\"end\":2.95}]"} {"utterance_id": "237-134500-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0012.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "When the utterance is complete, report all instances of a plant name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.73, 2.24, "cherries"]], "response_mode": "offline", "answer": "[{\"word\":\"cherries\",\"start\":1.73,\"end\":2.24}]"} {"utterance_id": "121-127105-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0020.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.2, 10.71, "douglas"]], "response_mode": "offline", "answer": "[{\"word\":\"douglas\",\"start\":10.2,\"end\":10.71}]"} {"utterance_id": "7729-102255-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0037.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Do not answer until the audio ends. Then list each instance of an organization name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[9.61, 10.57, "free state hotel"]], "response_mode": "offline", "answer": "[{\"word\":\"free state hotel\",\"start\":9.61,\"end\":10.57}]"} {"utterance_id": "237-126133-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0002.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.44, 5.63, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.44,\"end\":5.63}]"} {"utterance_id": "1284-1180-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0031.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the full audio. After it finishes, find every instance of a color name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.05, 3.57, "green"]], "response_mode": "offline", "answer": "[{\"word\":\"green\",\"start\":3.05,\"end\":3.57}]"} {"utterance_id": "1580-141084-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0029.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[1.22, 1.53, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":1.22,\"end\":1.53}]"} {"utterance_id": "6930-75918-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0001.flac", "task_type": "lexical_detection", "query": "provision", "prompt": "After the audio ends, extract all occurrences of the word \"provision\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.88, 5.47, "provision"]], "response_mode": "offline", "answer": "[{\"word\":\"provision\",\"start\":4.88,\"end\":5.47}]"} {"utterance_id": "8463-287645-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0005.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an organization name with fields word, start, end (seconds). If none, reply [].", "labels": [[11.64, 12.27, "infirmary"]], "response_mode": "offline", "answer": "[{\"word\":\"infirmary\",\"start\":11.64,\"end\":12.27}]"} {"utterance_id": "4970-29095-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0002.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the full audio. After it finishes, find every instance of a family relation. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.05, 1.63, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":1.05,\"end\":1.63}]"} {"utterance_id": "2961-960-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0020.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a person's name (word, start, end). If none, reply [].", "labels": [[5.11, 5.67, "one"], [8.63, 9.55, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.11,\"end\":5.67},{\"word\":\"two\",\"start\":8.63,\"end\":9.55}]"} {"utterance_id": "1580-141083-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0023.flac", "task_type": "lexical_detection", "query": "dry", "prompt": "When the utterance is complete, report all instances of the word \"dry\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.27, 2.55, "dry"]], "response_mode": "offline", "answer": "[{\"word\":\"dry\",\"start\":2.27,\"end\":2.55}]"} {"utterance_id": "7176-88083-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0012.flac", "task_type": "lexical_detection", "query": "branch", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"branch\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.59, 2.0, "branch"]], "response_mode": "offline", "answer": "[{\"word\":\"branch\",\"start\":1.59,\"end\":2.0}]"} {"utterance_id": "4992-41797-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0019.flac", "task_type": "lexical_detection", "query": "nancy's", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"nancy's\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.62, 3.08, "nancy's"]], "response_mode": "offline", "answer": "[{\"word\":\"nancy's\",\"start\":2.62,\"end\":3.08}]"} {"utterance_id": "4970-29095-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0017.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Process the full audio offline. Find every instance of a city name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.16, 3.65, "city"]], "response_mode": "offline", "answer": "[{\"word\":\"city\",\"start\":3.16,\"end\":3.65}]"} {"utterance_id": "8455-210777-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0033.flac", "task_type": "lexical_detection", "query": "yielding", "prompt": "After the audio ends, extract all occurrences of the word \"yielding\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.36, 6.74, "yielding"]], "response_mode": "offline", "answer": "[{\"word\":\"yielding\",\"start\":6.36,\"end\":6.74}]"} {"utterance_id": "237-126133-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0016.flac", "task_type": "lexical_detection", "query": "self", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"self\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.23, 3.73, "self"]], "response_mode": "offline", "answer": "[{\"word\":\"self\",\"start\":3.23,\"end\":3.73}]"} {"utterance_id": "4507-16021-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0032.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.96, 9.58, "judge"]], "response_mode": "offline", "answer": "[{\"word\":\"judge\",\"start\":8.96,\"end\":9.58}]"} {"utterance_id": "1188-133604-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0012.flac", "task_type": "lexical_detection", "query": "colorist", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"colorist\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.51, 2.36, "colorist"]], "response_mode": "offline", "answer": "[{\"word\":\"colorist\",\"start\":1.51,\"end\":2.36}]"} {"utterance_id": "1320-122612-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0010.flac", "task_type": "lexical_detection", "query": "pointing", "prompt": "After the audio ends, extract all occurrences of the word \"pointing\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.85, 2.27, "pointing"]], "response_mode": "offline", "answer": "[{\"word\":\"pointing\",\"start\":1.85,\"end\":2.27}]"} {"utterance_id": "121-123852-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123852/121-123852-0004.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 1.02, "heart"], [7.62, 8.2, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":0.53,\"end\":1.02},{\"word\":\"eyes\",\"start\":7.62,\"end\":8.2}]"} {"utterance_id": "8455-210777-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0009.flac", "task_type": "lexical_detection", "query": "process", "prompt": "After the audio ends, extract all occurrences of the word \"process\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.71, 2.16, "process"]], "response_mode": "offline", "answer": "[{\"word\":\"process\",\"start\":1.71,\"end\":2.16}]"} {"utterance_id": "5105-28240-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0008.flac", "task_type": "lexical_detection", "query": "channel", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"channel\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.81, 1.36, "channel"]], "response_mode": "offline", "answer": "[{\"word\":\"channel\",\"start\":0.81,\"end\":1.36}]"} {"utterance_id": "8224-274384-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0010.flac", "task_type": "semantic_counting", "query": "organization", "prompt": "Listen to the audio in full, then emit a JSON array of count events for an organization name (word, start, end). If none, reply [].", "labels": [[2.84, 3.81, "one"], [5.87, 6.51, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.84,\"end\":3.81},{\"word\":\"two\",\"start\":5.87,\"end\":6.51}]"} {"utterance_id": "1995-1826-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0000.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an organization name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.56, 2.22, "societies"]], "response_mode": "offline", "answer": "[{\"word\":\"societies\",\"start\":1.56,\"end\":2.22}]"} {"utterance_id": "8555-284447-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0000.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "After listening to the whole recording, output timed count words for a profession (one, two, three, ...) as JSON. If none, reply [].", "labels": [[4.34, 4.82, "one"], [7.48, 7.95, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.34,\"end\":4.82},{\"word\":\"two\",\"start\":7.48,\"end\":7.95}]"} {"utterance_id": "1995-1837-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0007.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Once the audio has finished, identify every instance of a time of day. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.39, 1.9, "midday"]], "response_mode": "offline", "answer": "[{\"word\":\"midday\",\"start\":1.39,\"end\":1.9}]"} {"utterance_id": "8455-210777-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0021.flac", "task_type": "lexical_detection", "query": "long", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"long\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.89, 2.29, "long"]], "response_mode": "offline", "answer": "[{\"word\":\"long\",\"start\":1.89,\"end\":2.29}]"} {"utterance_id": "5142-33396-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0053.flac", "task_type": "lexical_detection", "query": "treasure", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"treasure\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.2, 2.55, "treasure"]], "response_mode": "offline", "answer": "[{\"word\":\"treasure\",\"start\":2.2,\"end\":2.55}]"} {"utterance_id": "4992-41797-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.28, 2.49, "alexander carey"]], "response_mode": "offline", "answer": "[{\"word\":\"alexander carey\",\"start\":1.28,\"end\":2.49}]"} {"utterance_id": "237-134500-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0024.flac", "task_type": "lexical_detection", "query": "york", "prompt": "After listening to the whole recording, output every instance of the word \"york\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.09, 2.52, "york"]], "response_mode": "offline", "answer": "[{\"word\":\"york\",\"start\":2.09,\"end\":2.52}]"} {"utterance_id": "1995-1826-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0018.flac", "task_type": "lexical_detection", "query": "regard", "prompt": "Process the full audio offline. Find every instance of the word \"regard\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.41, 0.81, "regard"]], "response_mode": "offline", "answer": "[{\"word\":\"regard\",\"start\":0.41,\"end\":0.81}]"} {"utterance_id": "8455-210777-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0010.flac", "task_type": "lexical_detection", "query": "neverbend", "prompt": "After the audio ends, extract all occurrences of the word \"neverbend\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.11, 2.73, "neverbend"]], "response_mode": "offline", "answer": "[{\"word\":\"neverbend\",\"start\":2.11,\"end\":2.73}]"} {"utterance_id": "8555-292519-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0014.flac", "task_type": "lexical_detection", "query": "alone", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"alone\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.04, 1.48, "alone"]], "response_mode": "offline", "answer": "[{\"word\":\"alone\",\"start\":1.04,\"end\":1.48}]"} {"utterance_id": "1580-141084-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0040.flac", "task_type": "lexical_detection", "query": "provided", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"provided\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.85, 3.38, "provided"]], "response_mode": "offline", "answer": "[{\"word\":\"provided\",\"start\":2.85,\"end\":3.38}]"} {"utterance_id": "260-123288-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0026.flac", "task_type": "lexical_detection", "query": "blown", "prompt": "Process the full audio offline. Find every instance of the word \"blown\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.89, "blown"]], "response_mode": "offline", "answer": "[{\"word\":\"blown\",\"start\":0.53,\"end\":0.89}]"} {"utterance_id": "1284-134647-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0006.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the full audio. After it finishes, find every instance of an organization name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.07, 9.89, "donatists"]], "response_mode": "offline", "answer": "[{\"word\":\"donatists\",\"start\":9.07,\"end\":9.89}]"} {"utterance_id": "6829-68769-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0000.flac", "task_type": "lexical_detection", "query": "questioned", "prompt": "After the audio ends, extract all occurrences of the word \"questioned\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.66, 9.08, "questioned"]], "response_mode": "offline", "answer": "[{\"word\":\"questioned\",\"start\":8.66,\"end\":9.08}]"} {"utterance_id": "7127-75947-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0019.flac", "task_type": "lexical_detection", "query": "dancing", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"dancing\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.07, 1.52, "dancing"]], "response_mode": "offline", "answer": "[{\"word\":\"dancing\",\"start\":1.07,\"end\":1.52}]"} {"utterance_id": "1580-141084-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0000.flac", "task_type": "lexical_detection", "query": "indian", "prompt": "When the utterance is complete, report all instances of the word \"indian\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.92, 1.44, "indian"]], "response_mode": "offline", "answer": "[{\"word\":\"indian\",\"start\":0.92,\"end\":1.44}]"} {"utterance_id": "8555-284447-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0007.flac", "task_type": "lexical_counting", "query": "paid", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"paid\" with word/start/end. If none, reply [].", "labels": [[1.56, 1.86, "one"], [4.63, 4.91, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.56,\"end\":1.86},{\"word\":\"two\",\"start\":4.63,\"end\":4.91}]"} {"utterance_id": "7021-79740-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0007.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the full audio. Count every occurrence of a person's name. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.61, 2.1, "one"], [6.21, 6.65, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.61,\"end\":2.1},{\"word\":\"two\",\"start\":6.21,\"end\":6.65}]"} {"utterance_id": "61-70968-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0017.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a state or province name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.99, 4.52, "cumberland's"]], "response_mode": "offline", "answer": "[{\"word\":\"cumberland's\",\"start\":3.99,\"end\":4.52}]"} {"utterance_id": "61-70968-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0050.flac", "task_type": "lexical_detection", "query": "fingers", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"fingers\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.73, 4.08, "fingers"]], "response_mode": "offline", "answer": "[{\"word\":\"fingers\",\"start\":3.73,\"end\":4.08}]"} {"utterance_id": "1188-133604-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0037.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.06, 9.27, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":9.06,\"end\":9.27}]"} {"utterance_id": "1089-134686-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0002.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a color name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.9, 2.16, "yellow"]], "response_mode": "offline", "answer": "[{\"word\":\"yellow\",\"start\":1.9,\"end\":2.16}]"} {"utterance_id": "1995-1836-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0003.flac", "task_type": "lexical_detection", "query": "intelligent", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"intelligent\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.92, 2.58, "intelligent"]], "response_mode": "offline", "answer": "[{\"word\":\"intelligent\",\"start\":1.92,\"end\":2.58}]"} {"utterance_id": "1221-135766-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0000.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "After listening to the whole recording, output every instance of a weather term with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.14, 9.82, "sunshine"]], "response_mode": "offline", "answer": "[{\"word\":\"sunshine\",\"start\":9.14,\"end\":9.82}]"} {"utterance_id": "6829-68769-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0017.flac", "task_type": "lexical_detection", "query": "jailer", "prompt": "After the audio ends, extract all occurrences of the word \"jailer\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.76, "jailer"]], "response_mode": "offline", "answer": "[{\"word\":\"jailer\",\"start\":2.35,\"end\":2.76}]"} {"utterance_id": "3729-6852-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0036.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Once the audio has finished, identify every instance of a city name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.32, 1.94, "paris"]], "response_mode": "offline", "answer": "[{\"word\":\"paris\",\"start\":1.32,\"end\":1.94}]"} {"utterance_id": "260-123286-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0012.flac", "task_type": "lexical_detection", "query": "fear", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"fear\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.6, 2.14, "fear"]], "response_mode": "offline", "answer": "[{\"word\":\"fear\",\"start\":1.6,\"end\":2.14}]"} {"utterance_id": "5639-40744-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0031.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[16.63, 17.19, "rodolfo"], [20.66, 21.12, "rodolfo"]], "response_mode": "offline", "answer": "[{\"word\":\"rodolfo\",\"start\":16.63,\"end\":17.19},{\"word\":\"rodolfo\",\"start\":20.66,\"end\":21.12}]"} {"utterance_id": "5683-32879-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0022.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.41, 1.98, "rachel"]], "response_mode": "offline", "answer": "[{\"word\":\"rachel\",\"start\":1.41,\"end\":1.98}]"} {"utterance_id": "2094-142345-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0059.flac", "task_type": "lexical_detection", "query": "blushing", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"blushing\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.61, 2.08, "blushing"]], "response_mode": "offline", "answer": "[{\"word\":\"blushing\",\"start\":1.61,\"end\":2.08}]"} {"utterance_id": "7127-75947-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0035.flac", "task_type": "lexical_detection", "query": "gracious", "prompt": "After the audio ends, extract all occurrences of the word \"gracious\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.48, 1.09, "gracious"]], "response_mode": "offline", "answer": "[{\"word\":\"gracious\",\"start\":0.48,\"end\":1.09}]"} {"utterance_id": "121-121726-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0013.flac", "task_type": "lexical_detection", "query": "woman", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"woman\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.45, 2.0, "woman"]], "response_mode": "offline", "answer": "[{\"word\":\"woman\",\"start\":1.45,\"end\":2.0}]"} {"utterance_id": "5105-28241-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0017.flac", "task_type": "lexical_detection", "query": "depth", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"depth\" with fields word, start, end (seconds). If none, reply [].", "labels": [[6.24, 6.68, "depth"]], "response_mode": "offline", "answer": "[{\"word\":\"depth\",\"start\":6.24,\"end\":6.68}]"} {"utterance_id": "3570-5695-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0012.flac", "task_type": "lexical_detection", "query": "repute", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"repute\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.82, 4.35, "repute"]], "response_mode": "offline", "answer": "[{\"word\":\"repute\",\"start\":3.82,\"end\":4.35}]"} {"utterance_id": "5142-33396-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0021.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Process the full audio offline. Find every instance of a money or currency term and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.12, 2.47, "wealth"]], "response_mode": "offline", "answer": "[{\"word\":\"wealth\",\"start\":2.12,\"end\":2.47}]"} {"utterance_id": "2961-960-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0019.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.4, 1.59, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.4,\"end\":1.59}]"} {"utterance_id": "3575-170457-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0020.flac", "task_type": "lexical_detection", "query": "beginning", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"beginning\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.37, 3.73, "beginning"]], "response_mode": "offline", "answer": "[{\"word\":\"beginning\",\"start\":3.37,\"end\":3.73}]"} {"utterance_id": "1284-134647-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0003.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.49, 1.22, "constantine"]], "response_mode": "offline", "answer": "[{\"word\":\"constantine\",\"start\":0.49,\"end\":1.22}]"} {"utterance_id": "1320-122617-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0036.flac", "task_type": "lexical_detection", "query": "bless", "prompt": "After listening to the whole recording, output every instance of the word \"bless\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.78, 1.09, "bless"]], "response_mode": "offline", "answer": "[{\"word\":\"bless\",\"start\":0.78,\"end\":1.09}]"} {"utterance_id": "260-123440-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0002.flac", "task_type": "semantic_counting", "query": "color", "prompt": "Once the audio has finished, list each occurrence of a color name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[0.8, 1.12, "one"], [3.75, 4.1, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.8,\"end\":1.12},{\"word\":\"two\",\"start\":3.75,\"end\":4.1}]"} {"utterance_id": "3575-170457-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0033.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[6.72, 7.12, "miss e"]], "response_mode": "offline", "answer": "[{\"word\":\"miss e\",\"start\":6.72,\"end\":7.12}]"} {"utterance_id": "6829-68771-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0036.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 0.83, "eliza"]], "response_mode": "offline", "answer": "[{\"word\":\"eliza\",\"start\":0.42,\"end\":0.83}]"} {"utterance_id": "1089-134686-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0031.flac", "task_type": "lexical_detection", "query": "fold", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"fold\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.69, 3.24, "fold"]], "response_mode": "offline", "answer": "[{\"word\":\"fold\",\"start\":2.69,\"end\":3.24}]"} {"utterance_id": "4507-16021-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0020.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.06, 7.32, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.06,\"end\":7.32}]"} {"utterance_id": "2094-142345-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0039.flac", "task_type": "lexical_detection", "query": "happen", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"happen\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.41, 2.7, "happen"]], "response_mode": "offline", "answer": "[{\"word\":\"happen\",\"start\":2.41,\"end\":2.7}]"} {"utterance_id": "61-70968-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0035.flac", "task_type": "lexical_detection", "query": "squire's", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"squire's\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.51, 1.96, "squire's"]], "response_mode": "offline", "answer": "[{\"word\":\"squire's\",\"start\":1.51,\"end\":1.96}]"} {"utterance_id": "4992-41806-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0008.flac", "task_type": "lexical_detection", "query": "rising", "prompt": "After the audio ends, extract all occurrences of the word \"rising\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.09, 4.66, "rising"]], "response_mode": "offline", "answer": "[{\"word\":\"rising\",\"start\":4.09,\"end\":4.66}]"} {"utterance_id": "5142-33396-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0022.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.18, 3.62, "viking"]], "response_mode": "offline", "answer": "[{\"word\":\"viking\",\"start\":3.18,\"end\":3.62}]"} {"utterance_id": "5142-36377-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0013.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.57, 3.87, "next"]], "response_mode": "offline", "answer": "[{\"word\":\"next\",\"start\":3.57,\"end\":3.87}]"} {"utterance_id": "672-122797-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0052.flac", "task_type": "lexical_detection", "query": "older", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"older\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.45, 2.98, "older"]], "response_mode": "offline", "answer": "[{\"word\":\"older\",\"start\":2.45,\"end\":2.98}]"} {"utterance_id": "1188-133604-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0034.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Once the audio has finished, identify every instance of an ordinal number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[18.05, 18.59, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":18.05,\"end\":18.59}]"} {"utterance_id": "260-123288-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0005.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "After listening to the whole recording, output every instance of a weather term with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.34, 2.89, "mist"]], "response_mode": "offline", "answer": "[{\"word\":\"mist\",\"start\":2.34,\"end\":2.89}]"} {"utterance_id": "672-122797-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0027.flac", "task_type": "lexical_detection", "query": "decorated", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"decorated\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.48, 4.12, "decorated"]], "response_mode": "offline", "answer": "[{\"word\":\"decorated\",\"start\":3.48,\"end\":4.12}]"} {"utterance_id": "2830-3980-0057", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0057.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.31, 1.67, "christ"], [6.2, 6.46, "god"]], "response_mode": "offline", "answer": "[{\"word\":\"christ\",\"start\":1.31,\"end\":1.67},{\"word\":\"god\",\"start\":6.2,\"end\":6.46}]"} {"utterance_id": "4992-41806-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.52, 0.87, "olive"], [3.37, 4.09, "mother carey"]], "response_mode": "offline", "answer": "[{\"word\":\"olive\",\"start\":0.52,\"end\":0.87},{\"word\":\"mother carey\",\"start\":3.37,\"end\":4.09}]"} {"utterance_id": "7176-88083-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0001.flac", "task_type": "lexical_detection", "query": "collar", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"collar\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.57, 5.08, "collar"]], "response_mode": "offline", "answer": "[{\"word\":\"collar\",\"start\":4.57,\"end\":5.08}]"} {"utterance_id": "3570-5694-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0005.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen to the full audio. After it finishes, find every instance of a food name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.65, 2.25, "victuals"], [4.06, 4.8, "beverages"]], "response_mode": "offline", "answer": "[{\"word\":\"victuals\",\"start\":1.65,\"end\":2.25},{\"word\":\"beverages\",\"start\":4.06,\"end\":4.8}]"} {"utterance_id": "8230-279154-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0011.flac", "task_type": "lexical_counting", "query": "feel", "prompt": "Process the full audio offline. For every occurrence of the word \"feel\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.99, 3.33, "one"], [5.16, 5.35, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.99,\"end\":3.33},{\"word\":\"two\",\"start\":5.16,\"end\":5.35}]"} {"utterance_id": "3575-170457-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0017.flac", "task_type": "lexical_detection", "query": "southey", "prompt": "After listening to the whole recording, output every instance of the word \"southey\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.72, 12.19, "southey"]], "response_mode": "offline", "answer": "[{\"word\":\"southey\",\"start\":11.72,\"end\":12.19}]"} {"utterance_id": "4446-2275-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0013.flac", "task_type": "lexical_detection", "query": "wish", "prompt": "After the audio ends, extract all occurrences of the word \"wish\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.43, 1.65, "wish"]], "response_mode": "offline", "answer": "[{\"word\":\"wish\",\"start\":1.43,\"end\":1.65}]"} {"utterance_id": "4992-23283-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.44, 1.11, "miss woodley"]], "response_mode": "offline", "answer": "[{\"word\":\"miss woodley\",\"start\":0.44,\"end\":1.11}]"} {"utterance_id": "2094-142345-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0001.flac", "task_type": "lexical_detection", "query": "gate", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"gate\" with fields word, start, end (seconds). If none, reply [].", "labels": [[5.49, 5.94, "gate"]], "response_mode": "offline", "answer": "[{\"word\":\"gate\",\"start\":5.49,\"end\":5.94}]"} {"utterance_id": "3570-5694-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0016.flac", "task_type": "lexical_counting", "query": "persons", "prompt": "Listen to the entire clip, then report the count sequence for the word \"persons\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.98, 6.48, "one"], [9.88, 10.33, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.98,\"end\":6.48},{\"word\":\"two\",\"start\":9.88,\"end\":10.33}]"} {"utterance_id": "7021-79740-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0001.flac", "task_type": "semantic_counting", "query": "kinship", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a family relation (word, start, end). If none, reply [].", "labels": [[2.01, 2.39, "one"], [3.78, 4.34, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.01,\"end\":2.39},{\"word\":\"two\",\"start\":3.78,\"end\":4.34}]"} {"utterance_id": "2961-961-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0000.flac", "task_type": "lexical_detection", "query": "begins", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"begins\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.32, 1.69, "begins"]], "response_mode": "offline", "answer": "[{\"word\":\"begins\",\"start\":1.32,\"end\":1.69}]"} {"utterance_id": "8463-294828-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0000.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.67, 1.22, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":0.67,\"end\":1.22}]"} {"utterance_id": "2961-961-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0011.flac", "task_type": "lexical_detection", "query": "annals", "prompt": "After listening to the whole recording, output every instance of the word \"annals\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.28, 4.81, "annals"]], "response_mode": "offline", "answer": "[{\"word\":\"annals\",\"start\":4.28,\"end\":4.81}]"} {"utterance_id": "1995-1826-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0024.flac", "task_type": "lexical_counting", "query": "fleece", "prompt": "Process the full audio offline. For every occurrence of the word \"fleece\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.87, 1.57, "one"], [2.85, 3.51, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.87,\"end\":1.57},{\"word\":\"two\",\"start\":2.85,\"end\":3.51}]"} {"utterance_id": "4446-2273-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0013.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Once the audio has finished, identify every instance of a city name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.72, 1.14, "paris"]], "response_mode": "offline", "answer": "[{\"word\":\"paris\",\"start\":0.72,\"end\":1.14}]"} {"utterance_id": "908-157963-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0021.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.13, 6.56, "worm"]], "response_mode": "offline", "answer": "[{\"word\":\"worm\",\"start\":6.13,\"end\":6.56}]"} {"utterance_id": "61-70970-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0037.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.43, 2.84, "ears"]], "response_mode": "offline", "answer": "[{\"word\":\"ears\",\"start\":2.43,\"end\":2.84}]"} {"utterance_id": "8455-210777-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0059.flac", "task_type": "lexical_detection", "query": "entering", "prompt": "Once the audio has finished, identify every instance of the word \"entering\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.41, 3.87, "entering"]], "response_mode": "offline", "answer": "[{\"word\":\"entering\",\"start\":3.41,\"end\":3.87}]"} {"utterance_id": "2961-961-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.81, 5.29, "solon"]], "response_mode": "offline", "answer": "[{\"word\":\"solon\",\"start\":4.81,\"end\":5.29}]"} {"utterance_id": "237-134500-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0038.flac", "task_type": "lexical_detection", "query": "anyhow", "prompt": "After listening to the whole recording, output every instance of the word \"anyhow\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.34, 0.69, "anyhow"]], "response_mode": "offline", "answer": "[{\"word\":\"anyhow\",\"start\":0.34,\"end\":0.69}]"} {"utterance_id": "4992-41806-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0014.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an ordinal number with fields word, start, end (seconds). If none, reply [].", "labels": [[10.76, 11.16, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":10.76,\"end\":11.16}]"} {"utterance_id": "237-134493-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0014.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Process the full audio offline. Find every instance of an organization name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.21, 2.67, "school"]], "response_mode": "offline", "answer": "[{\"word\":\"school\",\"start\":2.21,\"end\":2.67}]"} {"utterance_id": "1580-141084-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0026.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an organization name with fields word, start, end (seconds). If none, reply [].", "labels": [[6.16, 6.88, "court martial"]], "response_mode": "offline", "answer": "[{\"word\":\"court martial\",\"start\":6.16,\"end\":6.88}]"} {"utterance_id": "672-122797-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0004.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Listen to the full audio. After it finishes, find every instance of a weather term. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.23, 2.65, "snow"]], "response_mode": "offline", "answer": "[{\"word\":\"snow\",\"start\":2.23,\"end\":2.65}]"} {"utterance_id": "2300-131720-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0035.flac", "task_type": "lexical_counting", "query": "edison", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"edison\" with word/start/end. If none, reply [].", "labels": [[4.12, 4.72, "one"], [11.41, 11.84, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.12,\"end\":4.72},{\"word\":\"two\",\"start\":11.41,\"end\":11.84}]"} {"utterance_id": "1995-1836-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0014.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.5, 2.57, "mister vanderpool"]], "response_mode": "offline", "answer": "[{\"word\":\"mister vanderpool\",\"start\":1.5,\"end\":2.57}]"} {"utterance_id": "8555-284447-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0002.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a time of day with fields word, start, end (seconds). If none, reply [].", "labels": [[5.41, 5.93, "mornin"]], "response_mode": "offline", "answer": "[{\"word\":\"mornin\",\"start\":5.41,\"end\":5.93}]"} {"utterance_id": "3729-6852-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0040.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an ordinal number with fields word, start, end (seconds). If none, reply [].", "labels": [[0.85, 1.51, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":0.85,\"end\":1.51}]"} {"utterance_id": "8463-287645-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0007.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.41, 3.78, "half"], [5.93, 6.14, "one"], [19.35, 19.61, "few"]], "response_mode": "offline", "answer": "[{\"word\":\"half\",\"start\":3.41,\"end\":3.78},{\"word\":\"one\",\"start\":5.93,\"end\":6.14},{\"word\":\"few\",\"start\":19.35,\"end\":19.61}]"} {"utterance_id": "8555-284447-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0008.flac", "task_type": "semantic_detection", "query": "color", "prompt": "When the utterance is complete, report all instances of a color name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.37, 1.72, "blue"]], "response_mode": "offline", "answer": "[{\"word\":\"blue\",\"start\":1.37,\"end\":1.72}]"} {"utterance_id": "2094-142345-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0030.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Do not answer until the audio ends. Then list each instance of a color name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[14.37, 14.77, "white"]], "response_mode": "offline", "answer": "[{\"word\":\"white\",\"start\":14.37,\"end\":14.77}]"} {"utterance_id": "2094-142345-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0059.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.78, 1.38, "missus poyser"]], "response_mode": "offline", "answer": "[{\"word\":\"missus poyser\",\"start\":0.78,\"end\":1.38}]"} {"utterance_id": "1580-141083-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0033.flac", "task_type": "lexical_detection", "query": "thin", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"thin\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.6, 3.05, "thin"]], "response_mode": "offline", "answer": "[{\"word\":\"thin\",\"start\":2.6,\"end\":3.05}]"} {"utterance_id": "8455-210777-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0030.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.46, 1.48, "mister neverbend"]], "response_mode": "offline", "answer": "[{\"word\":\"mister neverbend\",\"start\":0.46,\"end\":1.48}]"} {"utterance_id": "2961-960-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0012.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.47, 0.89, "many"]], "response_mode": "offline", "answer": "[{\"word\":\"many\",\"start\":0.47,\"end\":0.89}]"} {"utterance_id": "4992-23283-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0006.flac", "task_type": "lexical_detection", "query": "mistaken", "prompt": "When the utterance is complete, report all instances of the word \"mistaken\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.22, 1.93, "mistaken"]], "response_mode": "offline", "answer": "[{\"word\":\"mistaken\",\"start\":1.22,\"end\":1.93}]"} {"utterance_id": "1188-133604-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0007.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.14, 13.74, "mandible"]], "response_mode": "offline", "answer": "[{\"word\":\"mandible\",\"start\":13.14,\"end\":13.74}]"} {"utterance_id": "6930-75918-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0003.flac", "task_type": "lexical_detection", "query": "treated", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"treated\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.22, 7.63, "treated"]], "response_mode": "offline", "answer": "[{\"word\":\"treated\",\"start\":7.22,\"end\":7.63}]"} {"utterance_id": "4970-29095-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0016.flac", "task_type": "lexical_detection", "query": "flushed", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"flushed\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.37, 4.94, "flushed"]], "response_mode": "offline", "answer": "[{\"word\":\"flushed\",\"start\":4.37,\"end\":4.94}]"} {"utterance_id": "4992-41797-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0002.flac", "task_type": "lexical_detection", "query": "grandfather", "prompt": "Once the audio has finished, identify every instance of the word \"grandfather\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.37, 1.08, "grandfather"]], "response_mode": "offline", "answer": "[{\"word\":\"grandfather\",\"start\":0.37,\"end\":1.08}]"} {"utterance_id": "1188-133604-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0044.flac", "task_type": "lexical_detection", "query": "join", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"join\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.23, 2.51, "join"]], "response_mode": "offline", "answer": "[{\"word\":\"join\",\"start\":2.23,\"end\":2.51}]"} {"utterance_id": "1580-141084-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0011.flac", "task_type": "lexical_detection", "query": "darkness", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"darkness\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.41, 1.89, "darkness"]], "response_mode": "offline", "answer": "[{\"word\":\"darkness\",\"start\":1.41,\"end\":1.89}]"} {"utterance_id": "1089-134691-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0014.flac", "task_type": "lexical_detection", "query": "harmonized", "prompt": "After the audio ends, extract all occurrences of the word \"harmonized\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.98, 3.74, "harmonized"]], "response_mode": "offline", "answer": "[{\"word\":\"harmonized\",\"start\":2.98,\"end\":3.74}]"} {"utterance_id": "1188-133604-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0015.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Once the audio has finished, identify every instance of a weather term. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.37, 5.74, "mist"]], "response_mode": "offline", "answer": "[{\"word\":\"mist\",\"start\":5.37,\"end\":5.74}]"} {"utterance_id": "5639-40744-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.33, 0.92, "rodolfo"]], "response_mode": "offline", "answer": "[{\"word\":\"rodolfo\",\"start\":0.33,\"end\":0.92}]"} {"utterance_id": "4446-2275-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0032.flac", "task_type": "lexical_detection", "query": "you've", "prompt": "Process the full audio offline. Find every instance of the word \"you've\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.58, 1.76, "you've"]], "response_mode": "offline", "answer": "[{\"word\":\"you've\",\"start\":1.58,\"end\":1.76}]"} {"utterance_id": "1580-141084-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0012.flac", "task_type": "lexical_detection", "query": "mouthed", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"mouthed\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.96, 1.32, "mouthed"]], "response_mode": "offline", "answer": "[{\"word\":\"mouthed\",\"start\":0.96,\"end\":1.32}]"} {"utterance_id": "237-134500-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0014.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Process the full audio offline. Find every instance of a plant name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.88, 3.73, "cherry trees"]], "response_mode": "offline", "answer": "[{\"word\":\"cherry trees\",\"start\":2.88,\"end\":3.73}]"} {"utterance_id": "2830-3980-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0003.flac", "task_type": "lexical_detection", "query": "beneath", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"beneath\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.29, 1.67, "beneath"]], "response_mode": "offline", "answer": "[{\"word\":\"beneath\",\"start\":1.29,\"end\":1.67}]"} {"utterance_id": "1188-133604-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0021.flac", "task_type": "lexical_detection", "query": "felicity", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"felicity\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.0, 4.46, "felicity"], [10.96, 11.58, "felicity"]], "response_mode": "offline", "answer": "[{\"word\":\"felicity\",\"start\":4.0,\"end\":4.46},{\"word\":\"felicity\",\"start\":10.96,\"end\":11.58}]"} {"utterance_id": "1580-141084-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0043.flac", "task_type": "lexical_detection", "query": "shoes", "prompt": "Once the audio has finished, identify every instance of the word \"shoes\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.74, 1.1, "shoes"]], "response_mode": "offline", "answer": "[{\"word\":\"shoes\",\"start\":0.74,\"end\":1.1}]"} {"utterance_id": "4446-2271-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0007.flac", "task_type": "lexical_detection", "query": "doesn't", "prompt": "After the audio ends, extract all occurrences of the word \"doesn't\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 0.68, "doesn't"]], "response_mode": "offline", "answer": "[{\"word\":\"doesn't\",\"start\":0.42,\"end\":0.68}]"} {"utterance_id": "260-123288-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0013.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a weather term with fields word, start, end (seconds). If none, reply [].", "labels": [[8.09, 8.61, "mists"], [10.27, 11.03, "whirlwind"]], "response_mode": "offline", "answer": "[{\"word\":\"mists\",\"start\":8.09,\"end\":8.61},{\"word\":\"whirlwind\",\"start\":10.27,\"end\":11.03}]"} {"utterance_id": "6829-68769-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0028.flac", "task_type": "semantic_detection", "query": "money", "prompt": "After listening to the whole recording, output every instance of a money or currency term with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.87, 2.31, "checks"]], "response_mode": "offline", "answer": "[{\"word\":\"checks\",\"start\":1.87,\"end\":2.31}]"} {"utterance_id": "7729-102255-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.3, 1.12, "sheriff jones"]], "response_mode": "offline", "answer": "[{\"word\":\"sheriff jones\",\"start\":0.3,\"end\":1.12}]"} {"utterance_id": "4446-2275-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0015.flac", "task_type": "lexical_detection", "query": "pulled", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"pulled\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.4, 0.65, "pulled"]], "response_mode": "offline", "answer": "[{\"word\":\"pulled\",\"start\":0.4,\"end\":0.65}]"} {"utterance_id": "1320-122612-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0012.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "Listen to the full audio. Count every occurrence of an animal name. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.22, 4.51, "one"], [9.38, 10.05, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.22,\"end\":4.51},{\"word\":\"two\",\"start\":9.38,\"end\":10.05}]"} {"utterance_id": "2830-3980-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0059.flac", "task_type": "lexical_detection", "query": "paul", "prompt": "After the audio ends, extract all occurrences of the word \"paul\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 0.78, "paul"]], "response_mode": "offline", "answer": "[{\"word\":\"paul\",\"start\":0.52,\"end\":0.78}]"} {"utterance_id": "4970-29095-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0035.flac", "task_type": "lexical_detection", "query": "useless", "prompt": "When the utterance is complete, report all instances of the word \"useless\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.32, 3.76, "useless"]], "response_mode": "offline", "answer": "[{\"word\":\"useless\",\"start\":3.32,\"end\":3.76}]"} {"utterance_id": "260-123286-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0003.flac", "task_type": "lexical_detection", "query": "anxious", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"anxious\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.95, 1.47, "anxious"], [6.3, 7.09, "anxious"]], "response_mode": "offline", "answer": "[{\"word\":\"anxious\",\"start\":0.95,\"end\":1.47},{\"word\":\"anxious\",\"start\":6.3,\"end\":7.09}]"} {"utterance_id": "5683-32879-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0007.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the full audio. After it finishes, find every instance of a color name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.92, 1.42, "pale"]], "response_mode": "offline", "answer": "[{\"word\":\"pale\",\"start\":0.92,\"end\":1.42}]"} {"utterance_id": "3729-6852-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0029.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.01, 3.24, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":3.01,\"end\":3.24}]"} {"utterance_id": "7021-79740-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0012.flac", "task_type": "lexical_detection", "query": "playthings", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"playthings\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.69, 2.41, "playthings"]], "response_mode": "offline", "answer": "[{\"word\":\"playthings\",\"start\":1.69,\"end\":2.41}]"} {"utterance_id": "121-121726-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0011.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 1.37, "husband"], [3.3, 3.83, "wife"]], "response_mode": "offline", "answer": "[{\"word\":\"husband\",\"start\":0.51,\"end\":1.37},{\"word\":\"wife\",\"start\":3.3,\"end\":3.83}]"} {"utterance_id": "237-134493-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.47, 3.33, "john bergson"]], "response_mode": "offline", "answer": "[{\"word\":\"john bergson\",\"start\":2.47,\"end\":3.33}]"} {"utterance_id": "7127-75946-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0027.flac", "task_type": "lexical_detection", "query": "madame", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"madame\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.21, 2.67, "madame"]], "response_mode": "offline", "answer": "[{\"word\":\"madame\",\"start\":2.21,\"end\":2.67}]"} {"utterance_id": "4446-2273-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0036.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.49, 0.99, "alexander"]], "response_mode": "offline", "answer": "[{\"word\":\"alexander\",\"start\":0.49,\"end\":0.99}]"} {"utterance_id": "6930-76324-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0013.flac", "task_type": "lexical_detection", "query": "anything", "prompt": "After listening to the whole recording, output every instance of the word \"anything\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.82, 1.4, "anything"]], "response_mode": "offline", "answer": "[{\"word\":\"anything\",\"start\":0.82,\"end\":1.4}]"} {"utterance_id": "672-122797-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0040.flac", "task_type": "lexical_detection", "query": "stood", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"stood\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.34, 2.69, "stood"]], "response_mode": "offline", "answer": "[{\"word\":\"stood\",\"start\":2.34,\"end\":2.69}]"} {"utterance_id": "1284-1181-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0005.flac", "task_type": "lexical_detection", "query": "exclaimed", "prompt": "Once the audio has finished, identify every instance of the word \"exclaimed\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.73, 3.26, "exclaimed"]], "response_mode": "offline", "answer": "[{\"word\":\"exclaimed\",\"start\":2.73,\"end\":3.26}]"} {"utterance_id": "1284-1181-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0018.flac", "task_type": "lexical_detection", "query": "asserted", "prompt": "Process the full audio offline. Find every instance of the word \"asserted\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.79, 2.29, "asserted"]], "response_mode": "offline", "answer": "[{\"word\":\"asserted\",\"start\":1.79,\"end\":2.29}]"} {"utterance_id": "7729-102255-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0043.flac", "task_type": "lexical_detection", "query": "shannon", "prompt": "When the utterance is complete, report all instances of the word \"shannon\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.95, 4.53, "shannon"]], "response_mode": "offline", "answer": "[{\"word\":\"shannon\",\"start\":3.95,\"end\":4.53}]"} {"utterance_id": "8555-284447-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0023.flac", "task_type": "lexical_counting", "query": "shiver", "prompt": "Process the full audio offline. For every occurrence of the word \"shiver\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.72, 1.05, "one"], [5.17, 5.54, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.72,\"end\":1.05},{\"word\":\"two\",\"start\":5.17,\"end\":5.54}]"} {"utterance_id": "8463-294825-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0000.flac", "task_type": "lexical_detection", "query": "beyond", "prompt": "Process the full audio offline. Find every instance of the word \"beyond\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.1, 1.51, "beyond"]], "response_mode": "offline", "answer": "[{\"word\":\"beyond\",\"start\":1.1,\"end\":1.51}]"} {"utterance_id": "8463-294825-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0011.flac", "task_type": "lexical_detection", "query": "punished", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"punished\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.29, 2.01, "punished"]], "response_mode": "offline", "answer": "[{\"word\":\"punished\",\"start\":1.29,\"end\":2.01}]"} {"utterance_id": "4077-13754-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0012.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Process the full audio offline. Find every instance of an organization name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.15, 4.55, "congress"], [15.73, 16.08, "court's"]], "response_mode": "offline", "answer": "[{\"word\":\"congress\",\"start\":4.15,\"end\":4.55},{\"word\":\"court's\",\"start\":15.73,\"end\":16.08}]"} {"utterance_id": "3575-170457-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0022.flac", "task_type": "lexical_detection", "query": "confess", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"confess\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.29, 1.91, "confess"]], "response_mode": "offline", "answer": "[{\"word\":\"confess\",\"start\":1.29,\"end\":1.91}]"} {"utterance_id": "6930-76324-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0027.flac", "task_type": "lexical_detection", "query": "weariness", "prompt": "Process the full audio offline. Find every instance of the word \"weariness\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.32, 1.93, "weariness"]], "response_mode": "offline", "answer": "[{\"word\":\"weariness\",\"start\":1.32,\"end\":1.93}]"} {"utterance_id": "3729-6852-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.0, 3.54, "silvia"]], "response_mode": "offline", "answer": "[{\"word\":\"silvia\",\"start\":3.0,\"end\":3.54}]"} {"utterance_id": "2094-142345-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0052.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "You will hear a complete utterance. When it ends, list every instance of a family relation you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.93, 2.49, "husband"]], "response_mode": "offline", "answer": "[{\"word\":\"husband\",\"start\":1.93,\"end\":2.49}]"} {"utterance_id": "4992-41797-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0009.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a state or province name with fields word, start, end (seconds). If none, reply [].", "labels": [[6.34, 6.95, "new england"]], "response_mode": "offline", "answer": "[{\"word\":\"new england\",\"start\":6.34,\"end\":6.95}]"} {"utterance_id": "1221-135767-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0022.flac", "task_type": "lexical_detection", "query": "soil", "prompt": "Process the full audio offline. Find every instance of the word \"soil\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.43, 7.88, "soil"]], "response_mode": "offline", "answer": "[{\"word\":\"soil\",\"start\":7.43,\"end\":7.88}]"} {"utterance_id": "4446-2275-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0040.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.88, 1.26, "bartley"]], "response_mode": "offline", "answer": "[{\"word\":\"bartley\",\"start\":0.88,\"end\":1.26}]"} {"utterance_id": "7176-88083-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0002.flac", "task_type": "lexical_detection", "query": "bright", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"bright\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.43, 6.86, "bright"]], "response_mode": "offline", "answer": "[{\"word\":\"bright\",\"start\":6.43,\"end\":6.86}]"} {"utterance_id": "5105-28233-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28233/5105-28233-0000.flac", "task_type": "lexical_detection", "query": "length", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"length\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.48, 0.78, "length"]], "response_mode": "offline", "answer": "[{\"word\":\"length\",\"start\":0.48,\"end\":0.78}]"} {"utterance_id": "3570-5695-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0000.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.13, 3.3, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":3.13,\"end\":3.3}]"} {"utterance_id": "2094-142345-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0040.flac", "task_type": "lexical_detection", "query": "preach", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"preach\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.63, 0.98, "preach"]], "response_mode": "offline", "answer": "[{\"word\":\"preach\",\"start\":0.63,\"end\":0.98}]"} {"utterance_id": "8555-292519-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0007.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Listen to the entire clip, then report the count sequence for a body part as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.88, 1.44, "one"], [6.65, 7.1, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.88,\"end\":1.44},{\"word\":\"two\",\"start\":6.65,\"end\":7.1}]"} {"utterance_id": "4077-13751-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0011.flac", "task_type": "lexical_detection", "query": "thousands", "prompt": "Once the audio has finished, identify every instance of the word \"thousands\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.65, 1.18, "thousands"]], "response_mode": "offline", "answer": "[{\"word\":\"thousands\",\"start\":0.65,\"end\":1.18}]"} {"utterance_id": "7021-85628-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0021.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.84, 4.29, "mother's"]], "response_mode": "offline", "answer": "[{\"word\":\"mother's\",\"start\":3.84,\"end\":4.29}]"} {"utterance_id": "237-126133-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0007.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "After listening to the whole recording, output timed count words for a person's name (one, two, three, ...) as JSON. If none, reply [].", "labels": [[0.71, 1.22, "one"], [2.63, 3.16, "two"], [7.08, 7.6, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.71,\"end\":1.22},{\"word\":\"two\",\"start\":2.63,\"end\":3.16},{\"word\":\"three\",\"start\":7.08,\"end\":7.6}]"} {"utterance_id": "260-123286-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0025.flac", "task_type": "lexical_detection", "query": "reptiles", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"reptiles\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.69, 3.33, "reptiles"]], "response_mode": "offline", "answer": "[{\"word\":\"reptiles\",\"start\":2.69,\"end\":3.33}]"} {"utterance_id": "6829-68769-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0030.flac", "task_type": "semantic_detection", "query": "money", "prompt": "You will hear a complete utterance. When it ends, list every instance of a money or currency term you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.94, 3.31, "check"]], "response_mode": "offline", "answer": "[{\"word\":\"check\",\"start\":2.94,\"end\":3.31}]"} {"utterance_id": "1188-133604-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0038.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.85, 4.2, "turner"]], "response_mode": "offline", "answer": "[{\"word\":\"turner\",\"start\":3.85,\"end\":4.2}]"} {"utterance_id": "1320-122617-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0020.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[0.4, 0.87, "scout"]], "response_mode": "offline", "answer": "[{\"word\":\"scout\",\"start\":0.4,\"end\":0.87}]"} {"utterance_id": "4077-13751-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0002.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.22, 1.57, "six"], [4.46, 4.96, "two score"], [7.97, 8.83, "hundred thousand"]], "response_mode": "offline", "answer": "[{\"word\":\"six\",\"start\":1.22,\"end\":1.57},{\"word\":\"two score\",\"start\":4.46,\"end\":4.96},{\"word\":\"hundred thousand\",\"start\":7.97,\"end\":8.83}]"} {"utterance_id": "2300-131720-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0028.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.19, 5.76, "inventors"]], "response_mode": "offline", "answer": "[{\"word\":\"inventors\",\"start\":5.19,\"end\":5.76}]"} {"utterance_id": "7176-88083-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0015.flac", "task_type": "lexical_detection", "query": "instantly", "prompt": "Process the full audio offline. Find every instance of the word \"instantly\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.59, 1.07, "instantly"]], "response_mode": "offline", "answer": "[{\"word\":\"instantly\",\"start\":0.59,\"end\":1.07}]"} {"utterance_id": "6930-81414-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0011.flac", "task_type": "lexical_detection", "query": "freedom", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"freedom\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.01, 1.61, "freedom"]], "response_mode": "offline", "answer": "[{\"word\":\"freedom\",\"start\":1.01,\"end\":1.61}]"} {"utterance_id": "7729-102255-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0004.flac", "task_type": "lexical_detection", "query": "convictions", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"convictions\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.21, 3.13, "convictions"]], "response_mode": "offline", "answer": "[{\"word\":\"convictions\",\"start\":2.21,\"end\":3.13}]"} {"utterance_id": "237-134500-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0039.flac", "task_type": "lexical_detection", "query": "used", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"used\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.71, 4.04, "used"]], "response_mode": "offline", "answer": "[{\"word\":\"used\",\"start\":3.71,\"end\":4.04}]"} {"utterance_id": "4446-2273-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0014.flac", "task_type": "lexical_detection", "query": "changes", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"changes\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.73, 1.27, "changes"]], "response_mode": "offline", "answer": "[{\"word\":\"changes\",\"start\":0.73,\"end\":1.27}]"} {"utterance_id": "2830-3980-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0039.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.73, 3.09, "paul"]], "response_mode": "offline", "answer": "[{\"word\":\"paul\",\"start\":2.73,\"end\":3.09}]"} {"utterance_id": "3729-6852-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0000.flac", "task_type": "semantic_counting", "query": "kinship", "prompt": "After the audio ends, tally every instance of a family relation. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[2.04, 2.58, "one"], [6.47, 7.09, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.04,\"end\":2.58},{\"word\":\"two\",\"start\":6.47,\"end\":7.09}]"} {"utterance_id": "8230-279154-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0033.flac", "task_type": "lexical_detection", "query": "illusion", "prompt": "After the audio ends, extract all occurrences of the word \"illusion\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.62, 7.31, "illusion"]], "response_mode": "offline", "answer": "[{\"word\":\"illusion\",\"start\":6.62,\"end\":7.31}]"} {"utterance_id": "1284-134647-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0002.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.3, 5.64, "emperor"]], "response_mode": "offline", "answer": "[{\"word\":\"emperor\",\"start\":5.3,\"end\":5.64}]"} {"utterance_id": "7176-88083-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0007.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.4, 3.86, "flukes"], [7.56, 7.97, "beak"], [9.4, 9.76, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"flukes\",\"start\":3.4,\"end\":3.86},{\"word\":\"beak\",\"start\":7.56,\"end\":7.97},{\"word\":\"eyes\",\"start\":9.4,\"end\":9.76}]"} {"utterance_id": "8555-292519-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0005.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "After the audio ends, extract all occurrences of a time of day. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.14, 8.59, "evening"]], "response_mode": "offline", "answer": "[{\"word\":\"evening\",\"start\":8.14,\"end\":8.59}]"} {"utterance_id": "3729-6852-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0000.flac", "task_type": "semantic_detection", "query": "food", "prompt": "When the utterance is complete, report all instances of a food name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.58, 4.92, "supper"]], "response_mode": "offline", "answer": "[{\"word\":\"supper\",\"start\":4.58,\"end\":4.92}]"} {"utterance_id": "2300-131720-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0001.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.62, 8.69, "three thousand"]], "response_mode": "offline", "answer": "[{\"word\":\"three thousand\",\"start\":7.62,\"end\":8.69}]"} {"utterance_id": "5639-40744-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0021.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.22, 2.79, "parents"]], "response_mode": "offline", "answer": "[{\"word\":\"parents\",\"start\":2.22,\"end\":2.79}]"} {"utterance_id": "8463-294828-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0034.flac", "task_type": "lexical_detection", "query": "comfortable", "prompt": "Process the full audio offline. Find every instance of the word \"comfortable\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.21, 1.68, "comfortable"]], "response_mode": "offline", "answer": "[{\"word\":\"comfortable\",\"start\":1.21,\"end\":1.68}]"} {"utterance_id": "4992-41806-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0010.flac", "task_type": "lexical_detection", "query": "greatest", "prompt": "After the audio ends, extract all occurrences of the word \"greatest\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.18, 1.96, "greatest"]], "response_mode": "offline", "answer": "[{\"word\":\"greatest\",\"start\":1.18,\"end\":1.96}]"} {"utterance_id": "7176-92135-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0003.flac", "task_type": "lexical_detection", "query": "merely", "prompt": "When the utterance is complete, report all instances of the word \"merely\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.24, 1.65, "merely"]], "response_mode": "offline", "answer": "[{\"word\":\"merely\",\"start\":1.24,\"end\":1.65}]"} {"utterance_id": "121-127105-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0010.flac", "task_type": "lexical_detection", "query": "sent", "prompt": "Process the full audio offline. Find every instance of the word \"sent\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.36, 0.65, "sent"]], "response_mode": "offline", "answer": "[{\"word\":\"sent\",\"start\":0.36,\"end\":0.65}]"} {"utterance_id": "7176-92135-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0005.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Process the full audio offline. Find every instance of an organization name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.92, 5.34, "punch"]], "response_mode": "offline", "answer": "[{\"word\":\"punch\",\"start\":4.92,\"end\":5.34}]"} {"utterance_id": "260-123286-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0026.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[3.91, 4.47, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":3.91,\"end\":4.47}]"} {"utterance_id": "1320-122617-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0005.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Once the audio has finished, identify every instance of an animal name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.63, 0.92, "bear"]], "response_mode": "offline", "answer": "[{\"word\":\"bear\",\"start\":0.63,\"end\":0.92}]"} {"utterance_id": "7127-75946-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0013.flac", "task_type": "lexical_detection", "query": "completed", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"completed\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.14, 1.68, "completed"]], "response_mode": "offline", "answer": "[{\"word\":\"completed\",\"start\":1.14,\"end\":1.68}]"} {"utterance_id": "5142-33396-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0030.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Once the audio has finished, identify every instance of a food name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.25, 2.55, "meat"]], "response_mode": "offline", "answer": "[{\"word\":\"meat\",\"start\":2.25,\"end\":2.55}]"} {"utterance_id": "8230-279154-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0019.flac", "task_type": "lexical_counting", "query": "feelings", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"feelings\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[6.2, 6.98, "one"], [11.82, 12.32, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.2,\"end\":6.98},{\"word\":\"two\",\"start\":11.82,\"end\":12.32}]"} {"utterance_id": "1995-1837-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0015.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "Listen to the entire clip, then report the count sequence for a plant name as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.75, 1.16, "one"], [3.71, 4.36, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.75,\"end\":1.16},{\"word\":\"two\",\"start\":3.71,\"end\":4.36}]"} {"utterance_id": "1188-133604-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0016.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[14.32, 14.75, "jeweler's"]], "response_mode": "offline", "answer": "[{\"word\":\"jeweler's\",\"start\":14.32,\"end\":14.75}]"} {"utterance_id": "7127-75947-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0008.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.52, 1.93, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":1.52,\"end\":1.93}]"} {"utterance_id": "7021-85628-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0007.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "When the utterance is complete, report all instances of an ordinal number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.31, 8.78, "third"]], "response_mode": "offline", "answer": "[{\"word\":\"third\",\"start\":8.31,\"end\":8.78}]"} {"utterance_id": "6829-68771-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0000.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Once the audio has finished, list each occurrence of a person's name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[4.1, 4.83, "one"], [6.11, 6.61, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.1,\"end\":4.83},{\"word\":\"two\",\"start\":6.11,\"end\":6.61}]"} {"utterance_id": "3729-6852-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0023.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.39, 7.8, "writer"]], "response_mode": "offline", "answer": "[{\"word\":\"writer\",\"start\":7.39,\"end\":7.8}]"} {"utterance_id": "61-70968-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0051.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.45, 1.91, "sheriff"]], "response_mode": "offline", "answer": "[{\"word\":\"sheriff\",\"start\":1.45,\"end\":1.91}]"} {"utterance_id": "7729-102255-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0002.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "After the audio ends, extract all occurrences of a state or province name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.03, 4.58, "states"]], "response_mode": "offline", "answer": "[{\"word\":\"states\",\"start\":4.03,\"end\":4.58}]"} {"utterance_id": "6930-81414-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0014.flac", "task_type": "lexical_detection", "query": "moon", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"moon\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.12, 1.68, "moon"]], "response_mode": "offline", "answer": "[{\"word\":\"moon\",\"start\":1.12,\"end\":1.68}]"} {"utterance_id": "7021-85628-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0011.flac", "task_type": "lexical_detection", "query": "princess", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"princess\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 0.92, "princess"]], "response_mode": "offline", "answer": "[{\"word\":\"princess\",\"start\":0.52,\"end\":0.92}]"} {"utterance_id": "6829-68769-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0048.flac", "task_type": "semantic_detection", "query": "money", "prompt": "When the utterance is complete, report all instances of a money or currency term including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.68, 0.99, "check"], [4.19, 4.51, "check"]], "response_mode": "offline", "answer": "[{\"word\":\"check\",\"start\":0.68,\"end\":0.99},{\"word\":\"check\",\"start\":4.19,\"end\":4.51}]"} {"utterance_id": "2830-3980-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0051.flac", "task_type": "lexical_detection", "query": "justification", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"justification\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.23, 4.11, "justification"]], "response_mode": "offline", "answer": "[{\"word\":\"justification\",\"start\":3.23,\"end\":4.11}]"} {"utterance_id": "8463-294825-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0009.flac", "task_type": "lexical_detection", "query": "pole", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"pole\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[19.25, 19.74, "pole"]], "response_mode": "offline", "answer": "[{\"word\":\"pole\",\"start\":19.25,\"end\":19.74}]"} {"utterance_id": "2830-3980-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0013.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.03, 1.53, "apostles"]], "response_mode": "offline", "answer": "[{\"word\":\"apostles\",\"start\":1.03,\"end\":1.53}]"} {"utterance_id": "61-70970-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0018.flac", "task_type": "lexical_detection", "query": "wearily", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"wearily\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.29, 1.73, "wearily"]], "response_mode": "offline", "answer": "[{\"word\":\"wearily\",\"start\":1.29,\"end\":1.73}]"} {"utterance_id": "1284-1181-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0021.flac", "task_type": "lexical_detection", "query": "scornful", "prompt": "Process the full audio offline. Find every instance of the word \"scornful\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.32, 1.91, "scornful"]], "response_mode": "offline", "answer": "[{\"word\":\"scornful\",\"start\":1.32,\"end\":1.91}]"} {"utterance_id": "8230-279154-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0013.flac", "task_type": "lexical_counting", "query": "parts", "prompt": "Once the audio has finished, list each occurrence of the word \"parts\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[3.87, 4.29, "one"], [11.14, 11.8, "two"], [13.65, 14.27, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.87,\"end\":4.29},{\"word\":\"two\",\"start\":11.14,\"end\":11.8},{\"word\":\"three\",\"start\":13.65,\"end\":14.27}]"} {"utterance_id": "260-123288-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0010.flac", "task_type": "lexical_detection", "query": "mast", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"mast\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.66, 1.11, "mast"]], "response_mode": "offline", "answer": "[{\"word\":\"mast\",\"start\":0.66,\"end\":1.11}]"} {"utterance_id": "5683-32866-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0015.flac", "task_type": "lexical_detection", "query": "altogether", "prompt": "Once the audio has finished, identify every instance of the word \"altogether\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.06, 1.71, "altogether"]], "response_mode": "offline", "answer": "[{\"word\":\"altogether\",\"start\":1.06,\"end\":1.71}]"} {"utterance_id": "5105-28240-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0022.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the full audio. After it finishes, find every instance of an ordinal number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.07, 1.46, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":1.07,\"end\":1.46}]"} {"utterance_id": "5142-33396-0065", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0065.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.76, 1.05, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":0.76,\"end\":1.05}]"} {"utterance_id": "4446-2273-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0025.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen carefully to the entire clip, then report every instance of a food name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.92, 4.46, "coffee"]], "response_mode": "offline", "answer": "[{\"word\":\"coffee\",\"start\":3.92,\"end\":4.46}]"} {"utterance_id": "1320-122617-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0011.flac", "task_type": "lexical_detection", "query": "lodge", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"lodge\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.62, 0.97, "lodge"]], "response_mode": "offline", "answer": "[{\"word\":\"lodge\",\"start\":0.62,\"end\":0.97}]"} {"utterance_id": "5683-32865-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0012.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[9.82, 10.18, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":9.82,\"end\":10.18}]"} {"utterance_id": "4992-41797-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0006.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.69, 3.07, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":2.69,\"end\":3.07}]"} {"utterance_id": "4507-16021-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0000.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.64, 1.09, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.64,\"end\":1.09}]"} {"utterance_id": "4970-29095-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0033.flac", "task_type": "lexical_counting", "query": "put", "prompt": "Listen to the full audio. Count every occurrence of the word \"put\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.74, 1.92, "one"], [4.25, 4.47, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.74,\"end\":1.92},{\"word\":\"two\",\"start\":4.25,\"end\":4.47}]"} {"utterance_id": "3729-6852-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0005.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Once the audio has finished, identify every instance of a country name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.35, 2.98, "france"]], "response_mode": "offline", "answer": "[{\"word\":\"france\",\"start\":2.35,\"end\":2.98}]"} {"utterance_id": "7176-92135-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0019.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.01, 1.52, "nine"]], "response_mode": "offline", "answer": "[{\"word\":\"nine\",\"start\":1.01,\"end\":1.52}]"} {"utterance_id": "2094-142345-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0012.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.77, 12.37, "painter"]], "response_mode": "offline", "answer": "[{\"word\":\"painter\",\"start\":11.77,\"end\":12.37}]"} {"utterance_id": "6930-75918-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0006.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[4.22, 4.53, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":4.22,\"end\":4.53}]"} {"utterance_id": "237-126133-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0001.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After the audio ends, extract all occurrences of a plant name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.92, 11.58, "flowers"]], "response_mode": "offline", "answer": "[{\"word\":\"flowers\",\"start\":10.92,\"end\":11.58}]"} {"utterance_id": "1580-141084-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0031.flac", "task_type": "lexical_detection", "query": "gilchrist", "prompt": "After the audio ends, extract all occurrences of the word \"gilchrist\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.34, 1.92, "gilchrist"]], "response_mode": "offline", "answer": "[{\"word\":\"gilchrist\",\"start\":1.34,\"end\":1.92}]"} {"utterance_id": "4446-2273-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0018.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a city name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.6, 3.21, "paris"]], "response_mode": "offline", "answer": "[{\"word\":\"paris\",\"start\":2.6,\"end\":3.21}]"} {"utterance_id": "237-134500-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0005.flac", "task_type": "lexical_detection", "query": "mowed", "prompt": "When the utterance is complete, report all instances of the word \"mowed\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.82, 2.27, "mowed"]], "response_mode": "offline", "answer": "[{\"word\":\"mowed\",\"start\":1.82,\"end\":2.27}]"} {"utterance_id": "7176-88083-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0024.flac", "task_type": "lexical_detection", "query": "luck", "prompt": "Process the full audio offline. Find every instance of the word \"luck\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.94, 2.14, "luck"]], "response_mode": "offline", "answer": "[{\"word\":\"luck\",\"start\":1.94,\"end\":2.14}]"} {"utterance_id": "2830-3980-0065", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0065.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Once the audio has finished, list each occurrence of a person's name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[0.32, 0.6, "one"], [2.57, 3.5, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.32,\"end\":0.6},{\"word\":\"two\",\"start\":2.57,\"end\":3.5}]"} {"utterance_id": "7127-75947-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0012.flac", "task_type": "lexical_detection", "query": "indeed", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"indeed\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.55, 1.23, "indeed"]], "response_mode": "offline", "answer": "[{\"word\":\"indeed\",\"start\":0.55,\"end\":1.23}]"} {"utterance_id": "1284-1181-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0000.flac", "task_type": "lexical_detection", "query": "examined", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"examined\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.98, 1.61, "examined"]], "response_mode": "offline", "answer": "[{\"word\":\"examined\",\"start\":0.98,\"end\":1.61}]"} {"utterance_id": "7729-102255-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0029.flac", "task_type": "lexical_detection", "query": "vengeance", "prompt": "After listening to the whole recording, output every instance of the word \"vengeance\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.17, 5.76, "vengeance"]], "response_mode": "offline", "answer": "[{\"word\":\"vengeance\",\"start\":5.17,\"end\":5.76}]"} {"utterance_id": "6829-68771-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.23, 2.74, "reynolds"], [7.37, 8.01, "hopkins"]], "response_mode": "offline", "answer": "[{\"word\":\"reynolds\",\"start\":2.23,\"end\":2.74},{\"word\":\"hopkins\",\"start\":7.37,\"end\":8.01}]"} {"utterance_id": "2961-961-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0020.flac", "task_type": "lexical_detection", "query": "ocean", "prompt": "Once the audio has finished, identify every instance of the word \"ocean\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.16, 5.66, "ocean"]], "response_mode": "offline", "answer": "[{\"word\":\"ocean\",\"start\":5.16,\"end\":5.66}]"} {"utterance_id": "5683-32866-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0005.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.79, 4.31, "shoulder"]], "response_mode": "offline", "answer": "[{\"word\":\"shoulder\",\"start\":3.79,\"end\":4.31}]"} {"utterance_id": "260-123286-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0010.flac", "task_type": "semantic_detection", "query": "date", "prompt": "Listen carefully to the entire clip, then report every instance of a date as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 2.27, "sunday august sixteenth"]], "response_mode": "offline", "answer": "[{\"word\":\"sunday august sixteenth\",\"start\":0.52,\"end\":2.27}]"} {"utterance_id": "3575-170457-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0008.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.77, 9.36, "clergyman's"]], "response_mode": "offline", "answer": "[{\"word\":\"clergyman's\",\"start\":8.77,\"end\":9.36}]"} {"utterance_id": "1284-134647-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0002.flac", "task_type": "lexical_detection", "query": "persecution", "prompt": "After listening to the whole recording, output every instance of the word \"persecution\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.8, 7.64, "persecution"]], "response_mode": "offline", "answer": "[{\"word\":\"persecution\",\"start\":6.8,\"end\":7.64}]"} {"utterance_id": "3575-170457-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0011.flac", "task_type": "lexical_detection", "query": "view", "prompt": "After the audio ends, extract all occurrences of the word \"view\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.7, 2.06, "view"]], "response_mode": "offline", "answer": "[{\"word\":\"view\",\"start\":1.7,\"end\":2.06}]"} {"utterance_id": "237-134500-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0006.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.16, 1.62, "roses"]], "response_mode": "offline", "answer": "[{\"word\":\"roses\",\"start\":1.16,\"end\":1.62}]"} {"utterance_id": "237-134493-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0006.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.81, 2.35, "athlete"]], "response_mode": "offline", "answer": "[{\"word\":\"athlete\",\"start\":1.81,\"end\":2.35}]"} {"utterance_id": "1580-141083-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0032.flac", "task_type": "lexical_detection", "query": "injustice", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"injustice\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.93, 2.63, "injustice"]], "response_mode": "offline", "answer": "[{\"word\":\"injustice\",\"start\":1.93,\"end\":2.63}]"} {"utterance_id": "5683-32879-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0019.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.82, 1.16, "few"]], "response_mode": "offline", "answer": "[{\"word\":\"few\",\"start\":0.82,\"end\":1.16}]"} {"utterance_id": "4446-2273-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0035.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.3, 0.63, "bartley"]], "response_mode": "offline", "answer": "[{\"word\":\"bartley\",\"start\":0.3,\"end\":0.63}]"} {"utterance_id": "5142-33396-0061", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0061.flac", "task_type": "lexical_detection", "query": "queen", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"queen\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.77, 1.08, "queen"]], "response_mode": "offline", "answer": "[{\"word\":\"queen\",\"start\":0.77,\"end\":1.08}]"} {"utterance_id": "2961-961-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0013.flac", "task_type": "lexical_detection", "query": "particulars", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"particulars\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.78, 4.79, "particulars"]], "response_mode": "offline", "answer": "[{\"word\":\"particulars\",\"start\":3.78,\"end\":4.79}]"} {"utterance_id": "8555-284447-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0020.flac", "task_type": "lexical_detection", "query": "attack", "prompt": "Process the full audio offline. Find every instance of the word \"attack\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.42, 3.91, "attack"]], "response_mode": "offline", "answer": "[{\"word\":\"attack\",\"start\":3.42,\"end\":3.91}]"} {"utterance_id": "7021-79740-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0011.flac", "task_type": "lexical_detection", "query": "route", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"route\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.79, 8.04, "route"]], "response_mode": "offline", "answer": "[{\"word\":\"route\",\"start\":7.79,\"end\":8.04}]"} {"utterance_id": "6829-68771-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0007.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Do not answer until the audio ends. Then list each instance of an organization name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.45, 4.72, "woman's political league"]], "response_mode": "offline", "answer": "[{\"word\":\"woman's political league\",\"start\":3.45,\"end\":4.72}]"} {"utterance_id": "4446-2275-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0016.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.47, 5.92, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":5.47,\"end\":5.92}]"} {"utterance_id": "2094-142345-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0043.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.18, 1.36, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":1.18,\"end\":1.36}]"} {"utterance_id": "7176-92135-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0025.flac", "task_type": "lexical_detection", "query": "keeping", "prompt": "Once the audio has finished, identify every instance of the word \"keeping\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[8.79, 9.19, "keeping"]], "response_mode": "offline", "answer": "[{\"word\":\"keeping\",\"start\":8.79,\"end\":9.19}]"} {"utterance_id": "908-31957-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0020.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.52, 3.04, "hearts"]], "response_mode": "offline", "answer": "[{\"word\":\"hearts\",\"start\":2.52,\"end\":3.04}]"} {"utterance_id": "237-126133-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0014.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a person's name with word/start/end. If none, reply [].", "labels": [[0.64, 1.32, "one"], [3.51, 3.94, "two"], [5.61, 6.09, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.64,\"end\":1.32},{\"word\":\"two\",\"start\":3.51,\"end\":3.94},{\"word\":\"three\",\"start\":5.61,\"end\":6.09}]"} {"utterance_id": "3570-5694-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0019.flac", "task_type": "lexical_detection", "query": "distinction", "prompt": "When the utterance is complete, report all instances of the word \"distinction\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.06, 1.61, "distinction"]], "response_mode": "offline", "answer": "[{\"word\":\"distinction\",\"start\":1.06,\"end\":1.61}]"} {"utterance_id": "4970-29095-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0015.flac", "task_type": "lexical_detection", "query": "get", "prompt": "When the utterance is complete, report all instances of the word \"get\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.02, 1.21, "get"]], "response_mode": "offline", "answer": "[{\"word\":\"get\",\"start\":1.02,\"end\":1.21}]"} {"utterance_id": "1580-141083-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0021.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.41, 1.65, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.41,\"end\":1.65}]"} {"utterance_id": "1089-134691-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0013.flac", "task_type": "lexical_detection", "query": "argue", "prompt": "After the audio ends, extract all occurrences of the word \"argue\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.4, 2.82, "argue"]], "response_mode": "offline", "answer": "[{\"word\":\"argue\",\"start\":2.4,\"end\":2.82}]"} {"utterance_id": "4970-29095-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0023.flac", "task_type": "lexical_detection", "query": "person", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"person\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.98, 3.58, "person"]], "response_mode": "offline", "answer": "[{\"word\":\"person\",\"start\":2.98,\"end\":3.58}]"} {"utterance_id": "61-70968-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0059.flac", "task_type": "lexical_detection", "query": "disappointment", "prompt": "After the audio ends, extract all occurrences of the word \"disappointment\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.78, 1.45, "disappointment"]], "response_mode": "offline", "answer": "[{\"word\":\"disappointment\",\"start\":0.78,\"end\":1.45}]"} {"utterance_id": "5105-28241-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0014.flac", "task_type": "lexical_detection", "query": "remarkable", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"remarkable\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.17, 2.73, "remarkable"]], "response_mode": "offline", "answer": "[{\"word\":\"remarkable\",\"start\":2.17,\"end\":2.73}]"} {"utterance_id": "1580-141083-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0032.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.43, 0.89, "watson"]], "response_mode": "offline", "answer": "[{\"word\":\"watson\",\"start\":0.43,\"end\":0.89}]"} {"utterance_id": "1320-122612-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0000.flac", "task_type": "lexical_detection", "query": "active", "prompt": "After the audio ends, extract all occurrences of the word \"active\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.21, 2.67, "active"]], "response_mode": "offline", "answer": "[{\"word\":\"active\",\"start\":2.21,\"end\":2.67}]"} {"utterance_id": "7176-88083-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0010.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.31, 3.84, "jaws"]], "response_mode": "offline", "answer": "[{\"word\":\"jaws\",\"start\":3.31,\"end\":3.84}]"} {"utterance_id": "1188-133604-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0009.flac", "task_type": "lexical_detection", "query": "precious", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"precious\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.61, 8.19, "precious"]], "response_mode": "offline", "answer": "[{\"word\":\"precious\",\"start\":7.61,\"end\":8.19}]"} {"utterance_id": "4446-2271-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0012.flac", "task_type": "lexical_detection", "query": "famously", "prompt": "After listening to the whole recording, output every instance of the word \"famously\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.02, 2.61, "famously"]], "response_mode": "offline", "answer": "[{\"word\":\"famously\",\"start\":2.02,\"end\":2.61}]"} {"utterance_id": "1995-1837-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0021.flac", "task_type": "lexical_detection", "query": "hope", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"hope\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.32, 0.65, "hope"]], "response_mode": "offline", "answer": "[{\"word\":\"hope\",\"start\":0.32,\"end\":0.65}]"} {"utterance_id": "260-123440-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.72, 3.2, "mabel"]], "response_mode": "offline", "answer": "[{\"word\":\"mabel\",\"start\":2.72,\"end\":3.2}]"} {"utterance_id": "237-134500-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0022.flac", "task_type": "lexical_detection", "query": "wondered", "prompt": "When the utterance is complete, report all instances of the word \"wondered\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.39, 2.74, "wondered"]], "response_mode": "offline", "answer": "[{\"word\":\"wondered\",\"start\":2.39,\"end\":2.74}]"} {"utterance_id": "7176-88083-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0013.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Process the full audio offline. Find every instance of an animal name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.01, 2.47, "teal"], [12.91, 13.32, "hawk"]], "response_mode": "offline", "answer": "[{\"word\":\"teal\",\"start\":2.01,\"end\":2.47},{\"word\":\"hawk\",\"start\":12.91,\"end\":13.32}]"} {"utterance_id": "2961-960-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.35, 14.88, "plato"]], "response_mode": "offline", "answer": "[{\"word\":\"plato\",\"start\":14.35,\"end\":14.88}]"} {"utterance_id": "672-122797-0067", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0067.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "You will hear a complete utterance. When it ends, list every instance of a plant name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.83, 2.13, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":1.83,\"end\":2.13}]"} {"utterance_id": "1580-141084-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0029.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[6.13, 6.77, "bannister"]], "response_mode": "offline", "answer": "[{\"word\":\"bannister\",\"start\":6.13,\"end\":6.77}]"} {"utterance_id": "6930-81414-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0007.flac", "task_type": "lexical_detection", "query": "attached", "prompt": "After the audio ends, extract all occurrences of the word \"attached\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.58, 4.25, "attached"]], "response_mode": "offline", "answer": "[{\"word\":\"attached\",\"start\":3.58,\"end\":4.25}]"} {"utterance_id": "908-157963-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0006.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "When the utterance is complete, report all instances of a time of day including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.68, 7.17, "evening"]], "response_mode": "offline", "answer": "[{\"word\":\"evening\",\"start\":6.68,\"end\":7.17}]"} {"utterance_id": "4507-16021-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0016.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.83, 2.72, "naturalist"]], "response_mode": "offline", "answer": "[{\"word\":\"naturalist\",\"start\":1.83,\"end\":2.72}]"} {"utterance_id": "8463-294828-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0028.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.1, 4.74, "conseil"]], "response_mode": "offline", "answer": "[{\"word\":\"conseil\",\"start\":4.1,\"end\":4.74}]"} {"utterance_id": "6829-68769-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0005.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Do not answer until the audio ends. Then list each instance of a money or currency term as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.96, 6.5, "cost"]], "response_mode": "offline", "answer": "[{\"word\":\"cost\",\"start\":5.96,\"end\":6.5}]"} {"utterance_id": "8463-294828-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0036.flac", "task_type": "semantic_detection", "query": "city", "prompt": "You will hear a complete utterance. When it ends, list every instance of a city name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.27, 1.78, "brooklyn"], [2.83, 3.3, "new york"]], "response_mode": "offline", "answer": "[{\"word\":\"brooklyn\",\"start\":1.27,\"end\":1.78},{\"word\":\"new york\",\"start\":2.83,\"end\":3.3}]"} {"utterance_id": "1580-141084-0049", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0049.flac", "task_type": "lexical_detection", "query": "simple", "prompt": "Once the audio has finished, identify every instance of the word \"simple\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.61, 1.04, "simple"]], "response_mode": "offline", "answer": "[{\"word\":\"simple\",\"start\":0.61,\"end\":1.04}]"} {"utterance_id": "5639-40744-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0003.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[7.89, 8.47, "rodolfo"], [21.63, 22.41, "leocadia"]], "response_mode": "offline", "answer": "[{\"word\":\"rodolfo\",\"start\":7.89,\"end\":8.47},{\"word\":\"leocadia\",\"start\":21.63,\"end\":22.41}]"} {"utterance_id": "4077-13751-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0009.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "You will hear a complete utterance. When it ends, list every instance of a state or province name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.9, 5.43, "ohio"]], "response_mode": "offline", "answer": "[{\"word\":\"ohio\",\"start\":4.9,\"end\":5.43}]"} {"utterance_id": "1580-141084-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0016.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[4.01, 4.32, "half"]], "response_mode": "offline", "answer": "[{\"word\":\"half\",\"start\":4.01,\"end\":4.32}]"} {"utterance_id": "4077-13751-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0015.flac", "task_type": "semantic_detection", "query": "date", "prompt": "When the utterance is complete, report all instances of a date including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.01, 6.57, "october"]], "response_mode": "offline", "answer": "[{\"word\":\"october\",\"start\":6.01,\"end\":6.57}]"} {"utterance_id": "2830-3980-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0050.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.03, 6.42, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":6.03,\"end\":6.42}]"} {"utterance_id": "61-70970-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0010.flac", "task_type": "lexical_detection", "query": "dismiss", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"dismiss\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.1, 0.56, "dismiss"]], "response_mode": "offline", "answer": "[{\"word\":\"dismiss\",\"start\":0.1,\"end\":0.56}]"} {"utterance_id": "260-123288-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0021.flac", "task_type": "lexical_detection", "query": "heads", "prompt": "When the utterance is complete, report all instances of the word \"heads\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.57, 2.19, "heads"]], "response_mode": "offline", "answer": "[{\"word\":\"heads\",\"start\":1.57,\"end\":2.19}]"} {"utterance_id": "8230-279154-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0004.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.84, 6.33, "five"]], "response_mode": "offline", "answer": "[{\"word\":\"five\",\"start\":5.84,\"end\":6.33}]"} {"utterance_id": "6829-68769-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0004.flac", "task_type": "lexical_detection", "query": "innocent", "prompt": "After the audio ends, extract all occurrences of the word \"innocent\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.37, 3.88, "innocent"]], "response_mode": "offline", "answer": "[{\"word\":\"innocent\",\"start\":3.37,\"end\":3.88}]"} {"utterance_id": "260-123286-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0022.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.5, 0.78, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":0.5,\"end\":0.78}]"} {"utterance_id": "260-123286-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0021.flac", "task_type": "lexical_detection", "query": "watch", "prompt": "Once the audio has finished, identify every instance of the word \"watch\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.81, 1.29, "watch"]], "response_mode": "offline", "answer": "[{\"word\":\"watch\",\"start\":0.81,\"end\":1.29}]"} {"utterance_id": "6829-68769-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0015.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.53, 5.75, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.53,\"end\":5.75}]"} {"utterance_id": "8230-279154-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0042.flac", "task_type": "lexical_counting", "query": "recognize", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"recognize\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[1.12, 1.87, "one"], [6.35, 7.07, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.12,\"end\":1.87},{\"word\":\"two\",\"start\":6.35,\"end\":7.07}]"} {"utterance_id": "237-126133-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.8, 1.13, "jap"], [2.46, 3.01, "percy"], [4.79, 5.27, "jasper's"]], "response_mode": "offline", "answer": "[{\"word\":\"jap\",\"start\":0.8,\"end\":1.13},{\"word\":\"percy\",\"start\":2.46,\"end\":3.01},{\"word\":\"jasper's\",\"start\":4.79,\"end\":5.27}]"} {"utterance_id": "2300-131720-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0030.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.84, 2.33, "edison"]], "response_mode": "offline", "answer": "[{\"word\":\"edison\",\"start\":1.84,\"end\":2.33}]"} {"utterance_id": "5639-40744-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0003.flac", "task_type": "lexical_detection", "query": "encourage", "prompt": "Once the audio has finished, identify every instance of the word \"encourage\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[13.13, 13.66, "encourage"]], "response_mode": "offline", "answer": "[{\"word\":\"encourage\",\"start\":13.13,\"end\":13.66}]"} {"utterance_id": "1580-141083-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0050.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.16, 2.9, "mister holmes"]], "response_mode": "offline", "answer": "[{\"word\":\"mister holmes\",\"start\":2.16,\"end\":2.9}]"} {"utterance_id": "4077-13751-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0018.flac", "task_type": "lexical_counting", "query": "read", "prompt": "After listening to the whole recording, output timed count words for the word \"read\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[1.74, 1.95, "one"], [18.13, 18.38, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.74,\"end\":1.95},{\"word\":\"two\",\"start\":18.13,\"end\":18.38}]"} {"utterance_id": "2961-961-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0013.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.53, 1.09, "solon"]], "response_mode": "offline", "answer": "[{\"word\":\"solon\",\"start\":0.53,\"end\":1.09}]"} {"utterance_id": "7176-88083-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0019.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.56, 3.97, "half"]], "response_mode": "offline", "answer": "[{\"word\":\"half\",\"start\":3.56,\"end\":3.97}]"} {"utterance_id": "1284-1180-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0017.flac", "task_type": "lexical_counting", "query": "blue", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"blue\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[2.67, 2.92, "one"], [4.25, 4.55, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.67,\"end\":2.92},{\"word\":\"two\",\"start\":4.25,\"end\":4.55}]"} {"utterance_id": "4992-23283-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0003.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.46, 2.94, "sandford"]], "response_mode": "offline", "answer": "[{\"word\":\"sandford\",\"start\":2.46,\"end\":2.94}]"} {"utterance_id": "121-127105-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0024.flac", "task_type": "lexical_detection", "query": "committed", "prompt": "Process the full audio offline. Find every instance of the word \"committed\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.68, 4.08, "committed"]], "response_mode": "offline", "answer": "[{\"word\":\"committed\",\"start\":3.68,\"end\":4.08}]"} {"utterance_id": "1995-1837-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0007.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After listening to the whole recording, output every instance of a color name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.84, 7.06, "blue"]], "response_mode": "offline", "answer": "[{\"word\":\"blue\",\"start\":6.84,\"end\":7.06}]"} {"utterance_id": "4970-29093-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0018.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.62, 9.32, "operator"]], "response_mode": "offline", "answer": "[{\"word\":\"operator\",\"start\":8.62,\"end\":9.32}]"} {"utterance_id": "908-157963-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0011.flac", "task_type": "lexical_detection", "query": "cow", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"cow\" with fields word, start, end (seconds). If none, reply [].", "labels": [[6.45, 7.08, "cow"]], "response_mode": "offline", "answer": "[{\"word\":\"cow\",\"start\":6.45,\"end\":7.08}]"} {"utterance_id": "8230-279154-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0000.flac", "task_type": "lexical_detection", "query": "lecture", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"lecture\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.12, 4.72, "lecture"]], "response_mode": "offline", "answer": "[{\"word\":\"lecture\",\"start\":4.12,\"end\":4.72}]"} {"utterance_id": "7729-102255-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0035.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, list every instance of an organization name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.32, 1.25, "military force"]], "response_mode": "offline", "answer": "[{\"word\":\"military force\",\"start\":0.32,\"end\":1.25}]"} {"utterance_id": "5142-33396-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0024.flac", "task_type": "lexical_detection", "query": "reaching", "prompt": "Process the full audio offline. Find every instance of the word \"reaching\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.5, 3.88, "reaching"]], "response_mode": "offline", "answer": "[{\"word\":\"reaching\",\"start\":3.5,\"end\":3.88}]"} {"utterance_id": "7127-75946-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0018.flac", "task_type": "lexical_detection", "query": "buoyant", "prompt": "Once the audio has finished, identify every instance of the word \"buoyant\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.73, 3.19, "buoyant"]], "response_mode": "offline", "answer": "[{\"word\":\"buoyant\",\"start\":2.73,\"end\":3.19}]"} {"utterance_id": "672-122797-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0046.flac", "task_type": "lexical_detection", "query": "doors", "prompt": "After listening to the whole recording, output every instance of the word \"doors\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.41, 3.28, "doors"]], "response_mode": "offline", "answer": "[{\"word\":\"doors\",\"start\":2.41,\"end\":3.28}]"} {"utterance_id": "672-122797-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0046.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a plant name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.84, 4.29, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":3.84,\"end\":4.29}]"} {"utterance_id": "6930-76324-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0023.flac", "task_type": "lexical_detection", "query": "pocket", "prompt": "When the utterance is complete, report all instances of the word \"pocket\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.43, 0.86, "pocket"]], "response_mode": "offline", "answer": "[{\"word\":\"pocket\",\"start\":0.43,\"end\":0.86}]"} {"utterance_id": "7021-79759-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79759/7021-79759-0005.flac", "task_type": "lexical_detection", "query": "hasty", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"hasty\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.11, 3.76, "hasty"]], "response_mode": "offline", "answer": "[{\"word\":\"hasty\",\"start\":3.11,\"end\":3.76}]"} {"utterance_id": "1580-141084-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0024.flac", "task_type": "lexical_detection", "query": "mental", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"mental\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.55, 2.86, "mental"]], "response_mode": "offline", "answer": "[{\"word\":\"mental\",\"start\":2.55,\"end\":2.86}]"} {"utterance_id": "7021-85628-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0027.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[3.3, 3.72, "toe"]], "response_mode": "offline", "answer": "[{\"word\":\"toe\",\"start\":3.3,\"end\":3.72}]"} {"utterance_id": "4970-29093-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0006.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.58, 10.96, "clerk"]], "response_mode": "offline", "answer": "[{\"word\":\"clerk\",\"start\":10.58,\"end\":10.96}]"} {"utterance_id": "4507-16021-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0020.flac", "task_type": "lexical_detection", "query": "stopped", "prompt": "Once the audio has finished, identify every instance of the word \"stopped\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.0, 1.8, "stopped"]], "response_mode": "offline", "answer": "[{\"word\":\"stopped\",\"start\":1.0,\"end\":1.8}]"} {"utterance_id": "121-121726-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0009.flac", "task_type": "lexical_detection", "query": "gives", "prompt": "Process the full audio offline. Find every instance of the word \"gives\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.05, 4.37, "gives"]], "response_mode": "offline", "answer": "[{\"word\":\"gives\",\"start\":4.05,\"end\":4.37}]"} {"utterance_id": "5142-36600-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36600/5142-36600-0001.flac", "task_type": "lexical_detection", "query": "naturalists", "prompt": "After listening to the whole recording, output every instance of the word \"naturalists\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.2, 5.85, "naturalists"]], "response_mode": "offline", "answer": "[{\"word\":\"naturalists\",\"start\":5.2,\"end\":5.85}]"} {"utterance_id": "6829-68771-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0024.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.78, 5.05, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":4.78,\"end\":5.05}]"} {"utterance_id": "1320-122617-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0027.flac", "task_type": "lexical_detection", "query": "instructions", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"instructions\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.62, 5.46, "instructions"]], "response_mode": "offline", "answer": "[{\"word\":\"instructions\",\"start\":4.62,\"end\":5.46}]"} {"utterance_id": "8455-210777-0061", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0061.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.44, 1.88, "one hundred"]], "response_mode": "offline", "answer": "[{\"word\":\"one hundred\",\"start\":1.44,\"end\":1.88}]"} {"utterance_id": "5683-32866-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0003.flac", "task_type": "lexical_detection", "query": "formed", "prompt": "Process the full audio offline. Find every instance of the word \"formed\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.3, 1.6, "formed"]], "response_mode": "offline", "answer": "[{\"word\":\"formed\",\"start\":1.3,\"end\":1.6}]"} {"utterance_id": "4077-13751-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0015.flac", "task_type": "semantic_detection", "query": "city", "prompt": "After listening to the whole recording, output every instance of a city name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.86, 1.39, "settlement"]], "response_mode": "offline", "answer": "[{\"word\":\"settlement\",\"start\":0.86,\"end\":1.39}]"} {"utterance_id": "4077-13751-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0012.flac", "task_type": "lexical_detection", "query": "boggs", "prompt": "When the utterance is complete, report all instances of the word \"boggs\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.52, 3.11, "boggs"]], "response_mode": "offline", "answer": "[{\"word\":\"boggs\",\"start\":2.52,\"end\":3.11}]"} {"utterance_id": "5142-33396-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0008.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.75, 1.06, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":0.75,\"end\":1.06}]"} {"utterance_id": "1995-1836-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0013.flac", "task_type": "lexical_detection", "query": "related", "prompt": "Once the audio has finished, identify every instance of the word \"related\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.41, 1.84, "related"]], "response_mode": "offline", "answer": "[{\"word\":\"related\",\"start\":1.41,\"end\":1.84}]"} {"utterance_id": "7021-79740-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0000.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.28, 3.78, "children"]], "response_mode": "offline", "answer": "[{\"word\":\"children\",\"start\":3.28,\"end\":3.78}]"} {"utterance_id": "61-70968-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0040.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.38, 3.75, "robin"]], "response_mode": "offline", "answer": "[{\"word\":\"robin\",\"start\":3.38,\"end\":3.75}]"} {"utterance_id": "8463-287645-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0013.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.46, 3.79, "master jacob's"]], "response_mode": "offline", "answer": "[{\"word\":\"master jacob's\",\"start\":2.46,\"end\":3.79}]"} {"utterance_id": "8230-279154-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0001.flac", "task_type": "lexical_counting", "query": "sensation", "prompt": "Listen to the full audio. Count every occurrence of the word \"sensation\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.9, 3.78, "one"], [16.32, 17.12, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.9,\"end\":3.78},{\"word\":\"two\",\"start\":16.32,\"end\":17.12}]"} {"utterance_id": "6930-81414-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0013.flac", "task_type": "lexical_detection", "query": "scarcely", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"scarcely\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.77, 1.38, "scarcely"]], "response_mode": "offline", "answer": "[{\"word\":\"scarcely\",\"start\":0.77,\"end\":1.38}]"} {"utterance_id": "5683-32879-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0019.flac", "task_type": "lexical_counting", "query": "love", "prompt": "Listen to the full audio. Count every occurrence of the word \"love\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.32, 1.54, "one"], [3.6, 3.85, "two"], [5.56, 5.85, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.32,\"end\":1.54},{\"word\":\"two\",\"start\":3.6,\"end\":3.85},{\"word\":\"three\",\"start\":5.56,\"end\":5.85}]"} {"utterance_id": "5639-40744-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0021.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "When the utterance is complete, report how a body part accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.81, 8.09, "one"], [10.07, 10.57, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.81,\"end\":8.09},{\"word\":\"two\",\"start\":10.07,\"end\":10.57}]"} {"utterance_id": "7729-102255-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0045.flac", "task_type": "lexical_detection", "query": "martin", "prompt": "After the audio ends, extract all occurrences of the word \"martin\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.68, 1.12, "martin"]], "response_mode": "offline", "answer": "[{\"word\":\"martin\",\"start\":0.68,\"end\":1.12}]"} {"utterance_id": "7176-92135-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0012.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "After the audio ends, tally every instance of a profession. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[4.39, 4.8, "one"], [15.17, 15.78, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.39,\"end\":4.8},{\"word\":\"two\",\"start\":15.17,\"end\":15.78}]"} {"utterance_id": "4446-2273-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0032.flac", "task_type": "lexical_detection", "query": "steady", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"steady\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.14, 2.45, "steady"]], "response_mode": "offline", "answer": "[{\"word\":\"steady\",\"start\":2.14,\"end\":2.45}]"} {"utterance_id": "7729-102255-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0016.flac", "task_type": "lexical_detection", "query": "legality", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"legality\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.94, 3.76, "legality"]], "response_mode": "offline", "answer": "[{\"word\":\"legality\",\"start\":2.94,\"end\":3.76}]"} {"utterance_id": "7127-75946-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0009.flac", "task_type": "lexical_detection", "query": "contrary", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"contrary\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.35, 3.05, "contrary"]], "response_mode": "offline", "answer": "[{\"word\":\"contrary\",\"start\":2.35,\"end\":3.05}]"} {"utterance_id": "5142-33396-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0004.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.9, 1.17, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":0.9,\"end\":1.17}]"} {"utterance_id": "6829-68771-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0018.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.17, 1.43, "beth"]], "response_mode": "offline", "answer": "[{\"word\":\"beth\",\"start\":1.17,\"end\":1.43}]"} {"utterance_id": "3729-6852-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0046.flac", "task_type": "lexical_counting", "query": "never", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"never\" (word, start, end). If none, reply [].", "labels": [[3.45, 3.73, "one"], [10.11, 10.44, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.45,\"end\":3.73},{\"word\":\"two\",\"start\":10.11,\"end\":10.44}]"} {"utterance_id": "1089-134686-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0032.flac", "task_type": "lexical_detection", "query": "apostle", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"apostle\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.3, 2.97, "apostle"]], "response_mode": "offline", "answer": "[{\"word\":\"apostle\",\"start\":2.3,\"end\":2.97}]"} {"utterance_id": "2094-142345-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0033.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[2.77, 3.07, "eleven"]], "response_mode": "offline", "answer": "[{\"word\":\"eleven\",\"start\":2.77,\"end\":3.07}]"} {"utterance_id": "5142-36377-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0001.flac", "task_type": "lexical_detection", "query": "melancholy", "prompt": "Process the full audio offline. Find every instance of the word \"melancholy\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.47, 3.07, "melancholy"]], "response_mode": "offline", "answer": "[{\"word\":\"melancholy\",\"start\":2.47,\"end\":3.07}]"} {"utterance_id": "1188-133604-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0023.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a color name with fields word, start, end (seconds). If none, reply [].", "labels": [[5.44, 6.0, "ruby"], [8.61, 9.16, "black"], [14.77, 15.24, "black"], [20.55, 20.84, "black"]], "response_mode": "offline", "answer": "[{\"word\":\"ruby\",\"start\":5.44,\"end\":6.0},{\"word\":\"black\",\"start\":8.61,\"end\":9.16},{\"word\":\"black\",\"start\":14.77,\"end\":15.24},{\"word\":\"black\",\"start\":20.55,\"end\":20.84}]"} {"utterance_id": "8463-287645-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0013.flac", "task_type": "lexical_detection", "query": "age", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"age\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.86, 1.27, "age"]], "response_mode": "offline", "answer": "[{\"word\":\"age\",\"start\":0.86,\"end\":1.27}]"} {"utterance_id": "260-123288-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0015.flac", "task_type": "lexical_detection", "query": "innumerable", "prompt": "When the utterance is complete, report all instances of the word \"innumerable\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[15.22, 15.77, "innumerable"]], "response_mode": "offline", "answer": "[{\"word\":\"innumerable\",\"start\":15.22,\"end\":15.77}]"} {"utterance_id": "4507-16021-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0003.flac", "task_type": "lexical_detection", "query": "theft", "prompt": "When the utterance is complete, report all instances of the word \"theft\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.95, 2.57, "theft"]], "response_mode": "offline", "answer": "[{\"word\":\"theft\",\"start\":1.95,\"end\":2.57}]"} {"utterance_id": "2300-131720-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0017.flac", "task_type": "lexical_detection", "query": "multiple", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"multiple\" with fields word, start, end (seconds). If none, reply [].", "labels": [[6.04, 6.63, "multiple"]], "response_mode": "offline", "answer": "[{\"word\":\"multiple\",\"start\":6.04,\"end\":6.63}]"} {"utterance_id": "5142-36377-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0016.flac", "task_type": "lexical_detection", "query": "proved", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"proved\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.27, 1.69, "proved"]], "response_mode": "offline", "answer": "[{\"word\":\"proved\",\"start\":1.27,\"end\":1.69}]"} {"utterance_id": "4970-29093-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0000.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After the audio ends, extract all occurrences of an organization name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.46, 2.55, "astor library"]], "response_mode": "offline", "answer": "[{\"word\":\"astor library\",\"start\":1.46,\"end\":2.55}]"} {"utterance_id": "3575-170457-0056", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0056.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.85, 1.45, "branwell"]], "response_mode": "offline", "answer": "[{\"word\":\"branwell\",\"start\":0.85,\"end\":1.45}]"} {"utterance_id": "4446-2275-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0037.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.38, 0.87, "bartley"]], "response_mode": "offline", "answer": "[{\"word\":\"bartley\",\"start\":0.38,\"end\":0.87}]"} {"utterance_id": "5683-32865-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0007.flac", "task_type": "lexical_detection", "query": "wylder", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"wylder\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.04, 2.64, "wylder"]], "response_mode": "offline", "answer": "[{\"word\":\"wylder\",\"start\":2.04,\"end\":2.64}]"} {"utterance_id": "4077-13751-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0017.flac", "task_type": "lexical_detection", "query": "gaze", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"gaze\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.78, 3.03, "gaze"]], "response_mode": "offline", "answer": "[{\"word\":\"gaze\",\"start\":2.78,\"end\":3.03}]"} {"utterance_id": "260-123440-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0003.flac", "task_type": "lexical_detection", "query": "savage", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"savage\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.05, 1.7, "savage"]], "response_mode": "offline", "answer": "[{\"word\":\"savage\",\"start\":1.05,\"end\":1.7}]"} {"utterance_id": "1580-141084-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0024.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.93, 5.32, "holmes"]], "response_mode": "offline", "answer": "[{\"word\":\"holmes\",\"start\":4.93,\"end\":5.32}]"} {"utterance_id": "672-122797-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0019.flac", "task_type": "lexical_detection", "query": "fresh", "prompt": "After listening to the whole recording, output every instance of the word \"fresh\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.3, 2.94, "fresh"]], "response_mode": "offline", "answer": "[{\"word\":\"fresh\",\"start\":2.3,\"end\":2.94}]"} {"utterance_id": "8455-210777-0062", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0062.flac", "task_type": "lexical_detection", "query": "start", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"start\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.19, 2.69, "start"]], "response_mode": "offline", "answer": "[{\"word\":\"start\",\"start\":2.19,\"end\":2.69}]"} {"utterance_id": "4077-13751-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0002.flac", "task_type": "lexical_detection", "query": "score", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"score\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.65, 4.96, "score"]], "response_mode": "offline", "answer": "[{\"word\":\"score\",\"start\":4.65,\"end\":4.96}]"} {"utterance_id": "7021-85628-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0007.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.8, 2.27, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":1.8,\"end\":2.27}]"} {"utterance_id": "2300-131720-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0014.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.51, 1.91, "leader"]], "response_mode": "offline", "answer": "[{\"word\":\"leader\",\"start\":1.51,\"end\":1.91}]"} {"utterance_id": "1995-1837-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0019.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.51, 2.74, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.51,\"end\":2.74}]"} {"utterance_id": "2830-3980-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0050.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.92, 1.34, "christ"]], "response_mode": "offline", "answer": "[{\"word\":\"christ\",\"start\":0.92,\"end\":1.34}]"} {"utterance_id": "4992-41797-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0020.flac", "task_type": "lexical_detection", "query": "chestnut", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"chestnut\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.23, 1.87, "chestnut"]], "response_mode": "offline", "answer": "[{\"word\":\"chestnut\",\"start\":1.23,\"end\":1.87}]"} {"utterance_id": "4077-13751-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0005.flac", "task_type": "lexical_detection", "query": "seed", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"seed\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.74, 2.14, "seed"]], "response_mode": "offline", "answer": "[{\"word\":\"seed\",\"start\":1.74,\"end\":2.14}]"} {"utterance_id": "8455-210777-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.81, 2.73, "missus neverbend"]], "response_mode": "offline", "answer": "[{\"word\":\"missus neverbend\",\"start\":1.81,\"end\":2.73}]"} {"utterance_id": "2830-3980-0073", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0073.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.42, 14.88, "christ"]], "response_mode": "offline", "answer": "[{\"word\":\"christ\",\"start\":14.42,\"end\":14.88}]"} {"utterance_id": "1284-1180-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0020.flac", "task_type": "lexical_detection", "query": "lot", "prompt": "When the utterance is complete, report all instances of the word \"lot\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.67, 0.92, "lot"]], "response_mode": "offline", "answer": "[{\"word\":\"lot\",\"start\":0.67,\"end\":0.92}]"} {"utterance_id": "672-122797-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0002.flac", "task_type": "lexical_detection", "query": "cottage", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"cottage\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.34, 7.86, "cottage"]], "response_mode": "offline", "answer": "[{\"word\":\"cottage\",\"start\":7.34,\"end\":7.86}]"} {"utterance_id": "5639-40744-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0031.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the entire clip, then report the count sequence for a person's name as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[16.63, 17.19, "one"], [20.66, 21.12, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":16.63,\"end\":17.19},{\"word\":\"two\",\"start\":20.66,\"end\":21.12}]"} {"utterance_id": "8463-287645-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0005.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "You will hear a complete utterance. When it ends, list every instance of a weather term you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.2, 13.72, "frost"]], "response_mode": "offline", "answer": "[{\"word\":\"frost\",\"start\":13.2,\"end\":13.72}]"} {"utterance_id": "6930-81414-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0025.flac", "task_type": "lexical_detection", "query": "terrible", "prompt": "After the audio ends, extract all occurrences of the word \"terrible\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.73, 2.44, "terrible"]], "response_mode": "offline", "answer": "[{\"word\":\"terrible\",\"start\":1.73,\"end\":2.44}]"} {"utterance_id": "4446-2271-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0010.flac", "task_type": "lexical_detection", "query": "who's", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"who's\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.63, 0.82, "who's"]], "response_mode": "offline", "answer": "[{\"word\":\"who's\",\"start\":0.63,\"end\":0.82}]"} {"utterance_id": "1320-122617-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0021.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.73, 3.27, "six"]], "response_mode": "offline", "answer": "[{\"word\":\"six\",\"start\":2.73,\"end\":3.27}]"} {"utterance_id": "260-123288-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0019.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "You will hear a complete utterance. When it ends, list every instance of a time of day you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.72, 1.04, "noon"]], "response_mode": "offline", "answer": "[{\"word\":\"noon\",\"start\":0.72,\"end\":1.04}]"} {"utterance_id": "4446-2273-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0036.flac", "task_type": "lexical_detection", "query": "sides", "prompt": "After listening to the whole recording, output every instance of the word \"sides\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.22, 2.82, "sides"]], "response_mode": "offline", "answer": "[{\"word\":\"sides\",\"start\":2.22,\"end\":2.82}]"} {"utterance_id": "1284-1180-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0010.flac", "task_type": "lexical_detection", "query": "knocked", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"knocked\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.72, 1.04, "knocked"]], "response_mode": "offline", "answer": "[{\"word\":\"knocked\",\"start\":0.72,\"end\":1.04}]"} {"utterance_id": "1995-1836-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0001.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Do not answer until the audio ends. Then list each instance of an ordinal number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.42, 0.81, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":0.42,\"end\":0.81}]"} {"utterance_id": "4446-2273-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0010.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Process the full audio offline. Find every instance of an animal name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.75, 2.35, "sole"], [6.5, 7.19, "ducklings"]], "response_mode": "offline", "answer": "[{\"word\":\"sole\",\"start\":1.75,\"end\":2.35},{\"word\":\"ducklings\",\"start\":6.5,\"end\":7.19}]"} {"utterance_id": "4507-16021-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0043.flac", "task_type": "lexical_detection", "query": "jail", "prompt": "Process the full audio offline. Find every instance of the word \"jail\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.09, 3.73, "jail"]], "response_mode": "offline", "answer": "[{\"word\":\"jail\",\"start\":3.09,\"end\":3.73}]"} {"utterance_id": "7021-85628-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0013.flac", "task_type": "lexical_detection", "query": "dearly", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"dearly\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.17, 4.53, "dearly"]], "response_mode": "offline", "answer": "[{\"word\":\"dearly\",\"start\":4.17,\"end\":4.53}]"} {"utterance_id": "1221-135767-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0014.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[4.34, 4.56, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":4.34,\"end\":4.56}]"} {"utterance_id": "2300-131720-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0021.flac", "task_type": "lexical_detection", "query": "conductors", "prompt": "Once the audio has finished, identify every instance of the word \"conductors\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.97, 1.71, "conductors"]], "response_mode": "offline", "answer": "[{\"word\":\"conductors\",\"start\":0.97,\"end\":1.71}]"} {"utterance_id": "1188-133604-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0029.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "After listening to the whole recording, output every instance of a weather term with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.3, 2.59, "cloud"]], "response_mode": "offline", "answer": "[{\"word\":\"cloud\",\"start\":2.3,\"end\":2.59}]"} {"utterance_id": "1580-141084-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0017.flac", "task_type": "lexical_detection", "query": "watson", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"watson\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.37, 1.84, "watson"]], "response_mode": "offline", "answer": "[{\"word\":\"watson\",\"start\":1.37,\"end\":1.84}]"} {"utterance_id": "908-31957-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0015.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the full audio. After it finishes, find every instance of an ordinal number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.58, 7.95, "third"]], "response_mode": "offline", "answer": "[{\"word\":\"third\",\"start\":7.58,\"end\":7.95}]"} {"utterance_id": "3570-5694-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0007.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.96, 4.32, "body"]], "response_mode": "offline", "answer": "[{\"word\":\"body\",\"start\":3.96,\"end\":4.32}]"} {"utterance_id": "260-123440-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0011.flac", "task_type": "lexical_detection", "query": "i'll", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"i'll\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.23, 3.5, "i'll"]], "response_mode": "offline", "answer": "[{\"word\":\"i'll\",\"start\":3.23,\"end\":3.5}]"} {"utterance_id": "5639-40744-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0033.flac", "task_type": "lexical_detection", "query": "maids", "prompt": "When the utterance is complete, report all instances of the word \"maids\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.82, 6.31, "maids"]], "response_mode": "offline", "answer": "[{\"word\":\"maids\",\"start\":5.82,\"end\":6.31}]"} {"utterance_id": "4446-2273-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0008.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.63, 4.11, "sisters"], [5.9, 6.22, "cousin"]], "response_mode": "offline", "answer": "[{\"word\":\"sisters\",\"start\":3.63,\"end\":4.11},{\"word\":\"cousin\",\"start\":5.9,\"end\":6.22}]"} {"utterance_id": "2830-3980-0069", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0069.flac", "task_type": "lexical_detection", "query": "sins", "prompt": "Once the audio has finished, identify every instance of the word \"sins\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.88, 5.4, "sins"]], "response_mode": "offline", "answer": "[{\"word\":\"sins\",\"start\":4.88,\"end\":5.4}]"} {"utterance_id": "5683-32865-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0013.flac", "task_type": "lexical_detection", "query": "life", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"life\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.49, 6.92, "life"]], "response_mode": "offline", "answer": "[{\"word\":\"life\",\"start\":6.49,\"end\":6.92}]"} {"utterance_id": "2830-3980-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.37, 1.88, "christ"]], "response_mode": "offline", "answer": "[{\"word\":\"christ\",\"start\":1.37,\"end\":1.88}]"} {"utterance_id": "5142-33396-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0002.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.87, 1.31, "warriors"]], "response_mode": "offline", "answer": "[{\"word\":\"warriors\",\"start\":0.87,\"end\":1.31}]"} {"utterance_id": "3575-170457-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0038.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.2, 2.76, "jehovah"]], "response_mode": "offline", "answer": "[{\"word\":\"jehovah\",\"start\":2.2,\"end\":2.76}]"} {"utterance_id": "8224-274384-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0006.flac", "task_type": "lexical_detection", "query": "psalm", "prompt": "When the utterance is complete, report all instances of the word \"psalm\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.16, 3.82, "psalm"]], "response_mode": "offline", "answer": "[{\"word\":\"psalm\",\"start\":3.16,\"end\":3.82}]"} {"utterance_id": "7021-79740-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0001.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.01, 2.39, "sister"], [3.78, 4.34, "cousin"]], "response_mode": "offline", "answer": "[{\"word\":\"sister\",\"start\":2.01,\"end\":2.39},{\"word\":\"cousin\",\"start\":3.78,\"end\":4.34}]"} {"utterance_id": "7176-92135-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0012.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.49, 2.9, "hamlet's"]], "response_mode": "offline", "answer": "[{\"word\":\"hamlet's\",\"start\":2.49,\"end\":2.9}]"} {"utterance_id": "908-157963-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0019.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.04, 10.52, "wings"]], "response_mode": "offline", "answer": "[{\"word\":\"wings\",\"start\":10.04,\"end\":10.52}]"} {"utterance_id": "4992-41797-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0003.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 1.42, "mister popham"]], "response_mode": "offline", "answer": "[{\"word\":\"mister popham\",\"start\":0.52,\"end\":1.42}]"} {"utterance_id": "1089-134686-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0016.flac", "task_type": "lexical_detection", "query": "catechism", "prompt": "After listening to the whole recording, output every instance of the word \"catechism\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.84, 2.48, "catechism"]], "response_mode": "offline", "answer": "[{\"word\":\"catechism\",\"start\":1.84,\"end\":2.48}]"} {"utterance_id": "1089-134686-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0037.flac", "task_type": "semantic_counting", "query": "color", "prompt": "Once the audio has finished, list each occurrence of a color name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[1.61, 1.92, "one"], [4.12, 4.55, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.61,\"end\":1.92},{\"word\":\"two\",\"start\":4.12,\"end\":4.55}]"} {"utterance_id": "5142-33396-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0041.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.71, 1.12, "guards"]], "response_mode": "offline", "answer": "[{\"word\":\"guards\",\"start\":0.71,\"end\":1.12}]"} {"utterance_id": "5639-40744-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0026.flac", "task_type": "lexical_detection", "query": "daily", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"daily\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[7.41, 7.71, "daily"]], "response_mode": "offline", "answer": "[{\"word\":\"daily\",\"start\":7.41,\"end\":7.71}]"} {"utterance_id": "4077-13751-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0013.flac", "task_type": "lexical_detection", "query": "human", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"human\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.12, 3.47, "human"]], "response_mode": "offline", "answer": "[{\"word\":\"human\",\"start\":3.12,\"end\":3.47}]"} {"utterance_id": "1284-134647-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0001.flac", "task_type": "lexical_detection", "query": "confirmed", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"confirmed\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.07, 4.6, "confirmed"]], "response_mode": "offline", "answer": "[{\"word\":\"confirmed\",\"start\":4.07,\"end\":4.6}]"} {"utterance_id": "1995-1836-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0008.flac", "task_type": "lexical_detection", "query": "englishman", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"englishman\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.36, 4.91, "englishman"]], "response_mode": "offline", "answer": "[{\"word\":\"englishman\",\"start\":4.36,\"end\":4.91}]"} {"utterance_id": "6930-75918-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0011.flac", "task_type": "lexical_detection", "query": "convinced", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"convinced\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.49, 1.16, "convinced"]], "response_mode": "offline", "answer": "[{\"word\":\"convinced\",\"start\":0.49,\"end\":1.16}]"} {"utterance_id": "8230-279154-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0039.flac", "task_type": "lexical_detection", "query": "another", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"another\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.22, 3.72, "another"]], "response_mode": "offline", "answer": "[{\"word\":\"another\",\"start\":3.22,\"end\":3.72}]"} {"utterance_id": "1188-133604-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0023.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.62, 1.26, "colorist"]], "response_mode": "offline", "answer": "[{\"word\":\"colorist\",\"start\":0.62,\"end\":1.26}]"} {"utterance_id": "1995-1836-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0006.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.1, 3.89, "mister cresswell"]], "response_mode": "offline", "answer": "[{\"word\":\"mister cresswell\",\"start\":3.1,\"end\":3.89}]"} {"utterance_id": "3575-170457-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0008.flac", "task_type": "lexical_detection", "query": "dispersed", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"dispersed\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.96, 2.48, "dispersed"]], "response_mode": "offline", "answer": "[{\"word\":\"dispersed\",\"start\":1.96,\"end\":2.48}]"} {"utterance_id": "5683-32865-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0017.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.38, 3.83, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":3.38,\"end\":3.83}]"} {"utterance_id": "1221-135766-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0011.flac", "task_type": "lexical_detection", "query": "flight", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"flight\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.56, 7.04, "flight"]], "response_mode": "offline", "answer": "[{\"word\":\"flight\",\"start\":6.56,\"end\":7.04}]"} {"utterance_id": "5683-32865-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0016.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.67, 2.87, "lorne brandon"]], "response_mode": "offline", "answer": "[{\"word\":\"lorne brandon\",\"start\":1.67,\"end\":2.87}]"} {"utterance_id": "7729-102255-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0006.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen carefully to the entire clip, then report every instance of a city name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.06, 2.68, "towns"], [10.6, 11.23, "washington"]], "response_mode": "offline", "answer": "[{\"word\":\"towns\",\"start\":2.06,\"end\":2.68},{\"word\":\"washington\",\"start\":10.6,\"end\":11.23}]"} {"utterance_id": "1995-1826-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0023.flac", "task_type": "lexical_detection", "query": "tops", "prompt": "After listening to the whole recording, output every instance of the word \"tops\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.67, 2.04, "tops"]], "response_mode": "offline", "answer": "[{\"word\":\"tops\",\"start\":1.67,\"end\":2.04}]"} {"utterance_id": "8455-210777-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0038.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.77, 4.36, "president"], [7.77, 8.37, "servants"]], "response_mode": "offline", "answer": "[{\"word\":\"president\",\"start\":3.77,\"end\":4.36},{\"word\":\"servants\",\"start\":7.77,\"end\":8.37}]"} {"utterance_id": "4970-29095-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0005.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.66, 4.45, "relations"]], "response_mode": "offline", "answer": "[{\"word\":\"relations\",\"start\":3.66,\"end\":4.45}]"} {"utterance_id": "4446-2271-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0020.flac", "task_type": "lexical_counting", "query": "something", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"something\" with word/start/end. If none, reply [].", "labels": [[3.17, 3.53, "one"], [5.55, 5.9, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.17,\"end\":3.53},{\"word\":\"two\",\"start\":5.55,\"end\":5.9}]"} {"utterance_id": "5142-36377-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0024.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a number (word, start, end). If none, reply [].", "labels": [[3.24, 3.55, "one"], [4.9, 5.1, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.24,\"end\":3.55},{\"word\":\"two\",\"start\":4.9,\"end\":5.1}]"} {"utterance_id": "61-70968-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0050.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "You will hear a complete utterance. When it ends, list every instance of a disease name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.51, 1.93, "condition"]], "response_mode": "offline", "answer": "[{\"word\":\"condition\",\"start\":1.51,\"end\":1.93}]"} {"utterance_id": "1221-135767-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0002.flac", "task_type": "lexical_detection", "query": "simplicity", "prompt": "After listening to the whole recording, output every instance of the word \"simplicity\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.98, 2.72, "simplicity"]], "response_mode": "offline", "answer": "[{\"word\":\"simplicity\",\"start\":1.98,\"end\":2.72}]"} {"utterance_id": "2300-131720-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0023.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "When the utterance is complete, report all instances of an organization name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.03, 5.5, "edison construction department"]], "response_mode": "offline", "answer": "[{\"word\":\"edison construction department\",\"start\":4.03,\"end\":5.5}]"} {"utterance_id": "5639-40744-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0015.flac", "task_type": "lexical_detection", "query": "respect", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"respect\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.52, 6.02, "respect"]], "response_mode": "offline", "answer": "[{\"word\":\"respect\",\"start\":5.52,\"end\":6.02}]"} {"utterance_id": "2830-3980-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0014.flac", "task_type": "lexical_detection", "query": "lightly", "prompt": "Process the full audio offline. Find every instance of the word \"lightly\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.78, 2.25, "lightly"]], "response_mode": "offline", "answer": "[{\"word\":\"lightly\",\"start\":1.78,\"end\":2.25}]"} {"utterance_id": "260-123440-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.55, 0.85, "alice"]], "response_mode": "offline", "answer": "[{\"word\":\"alice\",\"start\":0.55,\"end\":0.85}]"} {"utterance_id": "5683-32866-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0027.flac", "task_type": "lexical_detection", "query": "shining", "prompt": "Process the full audio offline. Find every instance of the word \"shining\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.74, 2.3, "shining"]], "response_mode": "offline", "answer": "[{\"word\":\"shining\",\"start\":1.74,\"end\":2.3}]"} {"utterance_id": "260-123440-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0002.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[3.37, 3.62, "pair"], [4.99, 5.17, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"pair\",\"start\":3.37,\"end\":3.62},{\"word\":\"one\",\"start\":4.99,\"end\":5.17}]"} {"utterance_id": "61-70968-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0014.flac", "task_type": "lexical_detection", "query": "whistle", "prompt": "After listening to the whole recording, output every instance of the word \"whistle\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.77, 6.08, "whistle"]], "response_mode": "offline", "answer": "[{\"word\":\"whistle\",\"start\":5.77,\"end\":6.08}]"} {"utterance_id": "260-123286-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0000.flac", "task_type": "semantic_detection", "query": "date", "prompt": "After listening to the whole recording, output every instance of a date with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 2.39, "saturday august fifteenth"]], "response_mode": "offline", "answer": "[{\"word\":\"saturday august fifteenth\",\"start\":0.53,\"end\":2.39}]"} {"utterance_id": "5142-33396-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0009.flac", "task_type": "lexical_detection", "query": "foes", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"foes\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.66, 3.23, "foes"]], "response_mode": "offline", "answer": "[{\"word\":\"foes\",\"start\":2.66,\"end\":3.23}]"} {"utterance_id": "237-134493-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0000.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.89, 1.48, "sixteen"]], "response_mode": "offline", "answer": "[{\"word\":\"sixteen\",\"start\":0.89,\"end\":1.48}]"} {"utterance_id": "5142-36377-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0010.flac", "task_type": "lexical_detection", "query": "change", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"change\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.18, 3.5, "change"]], "response_mode": "offline", "answer": "[{\"word\":\"change\",\"start\":3.18,\"end\":3.5}]"} {"utterance_id": "1580-141083-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0053.flac", "task_type": "lexical_detection", "query": "haven't", "prompt": "After the audio ends, extract all occurrences of the word \"haven't\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.65, 0.96, "haven't"]], "response_mode": "offline", "answer": "[{\"word\":\"haven't\",\"start\":0.65,\"end\":0.96}]"} {"utterance_id": "5142-33396-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0007.flac", "task_type": "lexical_detection", "query": "thrust", "prompt": "After listening to the whole recording, output every instance of the word \"thrust\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.08, 4.5, "thrust"]], "response_mode": "offline", "answer": "[{\"word\":\"thrust\",\"start\":4.08,\"end\":4.5}]"} {"utterance_id": "237-134500-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0014.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 0.89, "emil"]], "response_mode": "offline", "answer": "[{\"word\":\"emil\",\"start\":0.51,\"end\":0.89}]"} {"utterance_id": "7176-92135-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0016.flac", "task_type": "lexical_detection", "query": "attention", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"attention\" with fields word, start, end (seconds). If none, reply [].", "labels": [[6.34, 6.78, "attention"]], "response_mode": "offline", "answer": "[{\"word\":\"attention\",\"start\":6.34,\"end\":6.78}]"} {"utterance_id": "1284-1181-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0013.flac", "task_type": "lexical_counting", "query": "cleverness", "prompt": "Once the audio has finished, list each occurrence of the word \"cleverness\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[4.49, 4.99, "one"], [11.5, 12.22, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.49,\"end\":4.99},{\"word\":\"two\",\"start\":11.5,\"end\":12.22}]"} {"utterance_id": "4507-16021-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0030.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Listen to the full audio. After it finishes, find every instance of a country name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.39, 3.96, "nation"]], "response_mode": "offline", "answer": "[{\"word\":\"nation\",\"start\":3.39,\"end\":3.96}]"} {"utterance_id": "7021-79730-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0007.flac", "task_type": "lexical_detection", "query": "show", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"show\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.39, 3.68, "show"]], "response_mode": "offline", "answer": "[{\"word\":\"show\",\"start\":3.39,\"end\":3.68}]"} {"utterance_id": "1995-1826-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0009.flac", "task_type": "lexical_detection", "query": "negroes", "prompt": "Process the full audio offline. Find every instance of the word \"negroes\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.13, 3.85, "negroes"]], "response_mode": "offline", "answer": "[{\"word\":\"negroes\",\"start\":3.13,\"end\":3.85}]"} {"utterance_id": "237-126133-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0011.flac", "task_type": "lexical_detection", "query": "intense", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"intense\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.91, 3.42, "intense"]], "response_mode": "offline", "answer": "[{\"word\":\"intense\",\"start\":2.91,\"end\":3.42}]"} {"utterance_id": "237-134500-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0018.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "Process the full audio offline. For every occurrence of a plant name, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.37, 2.04, "one"], [4.36, 4.77, "two"], [6.72, 7.24, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.37,\"end\":2.04},{\"word\":\"two\",\"start\":4.36,\"end\":4.77},{\"word\":\"three\",\"start\":6.72,\"end\":7.24}]"} {"utterance_id": "7021-79740-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0002.flac", "task_type": "lexical_detection", "query": "minds", "prompt": "After the audio ends, extract all occurrences of the word \"minds\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.72, 6.13, "minds"]], "response_mode": "offline", "answer": "[{\"word\":\"minds\",\"start\":5.72,\"end\":6.13}]"} {"utterance_id": "2830-3980-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0025.flac", "task_type": "lexical_detection", "query": "won", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"won\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.82, 2.06, "won"]], "response_mode": "offline", "answer": "[{\"word\":\"won\",\"start\":1.82,\"end\":2.06}]"} {"utterance_id": "908-157963-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0019.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "When the utterance is complete, report all instances of a weather term including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.96, 14.51, "dew"]], "response_mode": "offline", "answer": "[{\"word\":\"dew\",\"start\":13.96,\"end\":14.51}]"} {"utterance_id": "4446-2275-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0043.flac", "task_type": "lexical_detection", "query": "mouth", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"mouth\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.54, 3.92, "mouth"]], "response_mode": "offline", "answer": "[{\"word\":\"mouth\",\"start\":3.54,\"end\":3.92}]"} {"utterance_id": "1320-122617-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.2, 1.71, "hawkeye"]], "response_mode": "offline", "answer": "[{\"word\":\"hawkeye\",\"start\":1.2,\"end\":1.71}]"} {"utterance_id": "237-134493-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.91, 1.3, "lou's"]], "response_mode": "offline", "answer": "[{\"word\":\"lou's\",\"start\":0.91,\"end\":1.3}]"} {"utterance_id": "2961-960-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0009.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "After the audio ends, tally every instance of an animal name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[3.3, 3.66, "one"], [10.64, 11.33, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.3,\"end\":3.66},{\"word\":\"two\",\"start\":10.64,\"end\":11.33}]"} {"utterance_id": "2094-142345-0058", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0058.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.69, 4.41, "missus poyser"]], "response_mode": "offline", "answer": "[{\"word\":\"missus poyser\",\"start\":3.69,\"end\":4.41}]"} {"utterance_id": "1221-135767-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0007.flac", "task_type": "lexical_detection", "query": "contrived", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"contrived\" with fields word, start, end (seconds). If none, reply [].", "labels": [[7.33, 7.97, "contrived"]], "response_mode": "offline", "answer": "[{\"word\":\"contrived\",\"start\":7.33,\"end\":7.97}]"} {"utterance_id": "1995-1837-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0011.flac", "task_type": "lexical_detection", "query": "sadly", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"sadly\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.65, 3.21, "sadly"]], "response_mode": "offline", "answer": "[{\"word\":\"sadly\",\"start\":2.65,\"end\":3.21}]"} {"utterance_id": "672-122797-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0048.flac", "task_type": "lexical_detection", "query": "lonely", "prompt": "When the utterance is complete, report all instances of the word \"lonely\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.48, 6.14, "lonely"]], "response_mode": "offline", "answer": "[{\"word\":\"lonely\",\"start\":5.48,\"end\":6.14}]"} {"utterance_id": "1580-141084-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0024.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.13, 6.52, "hands"]], "response_mode": "offline", "answer": "[{\"word\":\"hands\",\"start\":6.13,\"end\":6.52}]"} {"utterance_id": "1320-122617-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0010.flac", "task_type": "lexical_detection", "query": "hesitating", "prompt": "Process the full audio offline. Find every instance of the word \"hesitating\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.9, 3.67, "hesitating"]], "response_mode": "offline", "answer": "[{\"word\":\"hesitating\",\"start\":2.9,\"end\":3.67}]"} {"utterance_id": "4077-13751-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0002.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Once the audio has finished, list each occurrence of a number as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[1.22, 1.57, "one"], [4.46, 4.96, "two"], [7.97, 8.83, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.22,\"end\":1.57},{\"word\":\"two\",\"start\":4.46,\"end\":4.96},{\"word\":\"three\",\"start\":7.97,\"end\":8.83}]"} {"utterance_id": "3575-170457-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0047.flac", "task_type": "lexical_detection", "query": "tabby", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"tabby\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.39, 0.72, "tabby"]], "response_mode": "offline", "answer": "[{\"word\":\"tabby\",\"start\":0.39,\"end\":0.72}]"} {"utterance_id": "8463-294828-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0038.flac", "task_type": "lexical_detection", "query": "lincoln", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"lincoln\" with fields word, start, end (seconds). If none, reply [].", "labels": [[6.41, 6.81, "lincoln"]], "response_mode": "offline", "answer": "[{\"word\":\"lincoln\",\"start\":6.41,\"end\":6.81}]"} {"utterance_id": "672-122797-0072", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0072.flac", "task_type": "lexical_detection", "query": "heap", "prompt": "Once the audio has finished, identify every instance of the word \"heap\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.27, 6.65, "heap"]], "response_mode": "offline", "answer": "[{\"word\":\"heap\",\"start\":6.27,\"end\":6.65}]"} {"utterance_id": "672-122797-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0001.flac", "task_type": "lexical_detection", "query": "comrades", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"comrades\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.99, 12.71, "comrades"]], "response_mode": "offline", "answer": "[{\"word\":\"comrades\",\"start\":11.99,\"end\":12.71}]"} {"utterance_id": "8230-279154-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0021.flac", "task_type": "lexical_detection", "query": "occurrence", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"occurrence\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.9, 2.54, "occurrence"]], "response_mode": "offline", "answer": "[{\"word\":\"occurrence\",\"start\":1.9,\"end\":2.54}]"} {"utterance_id": "1580-141084-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0022.flac", "task_type": "lexical_detection", "query": "morning", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"morning\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.33, 1.72, "morning"]], "response_mode": "offline", "answer": "[{\"word\":\"morning\",\"start\":1.33,\"end\":1.72}]"} {"utterance_id": "4507-16021-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0017.flac", "task_type": "lexical_detection", "query": "refusing", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"refusing\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.52, 3.09, "refusing"]], "response_mode": "offline", "answer": "[{\"word\":\"refusing\",\"start\":2.52,\"end\":3.09}]"} {"utterance_id": "4446-2275-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0005.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.04, 1.49, "bones"]], "response_mode": "offline", "answer": "[{\"word\":\"bones\",\"start\":1.04,\"end\":1.49}]"} {"utterance_id": "6930-75918-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0017.flac", "task_type": "lexical_detection", "query": "conflict", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"conflict\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.25, 5.94, "conflict"]], "response_mode": "offline", "answer": "[{\"word\":\"conflict\",\"start\":5.25,\"end\":5.94}]"} {"utterance_id": "6829-68771-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0032.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.89, 1.38, "features"], [4.68, 5.1, "hands"]], "response_mode": "offline", "answer": "[{\"word\":\"features\",\"start\":0.89,\"end\":1.38},{\"word\":\"hands\",\"start\":4.68,\"end\":5.1}]"} {"utterance_id": "908-31957-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0006.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "After listening to the whole recording, output timed count words for a body part (one, two, three, ...) as JSON. If none, reply [].", "labels": [[0.88, 1.24, "one"], [4.11, 4.57, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.88,\"end\":1.24},{\"word\":\"two\",\"start\":4.11,\"end\":4.57}]"} {"utterance_id": "8555-292519-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0010.flac", "task_type": "lexical_detection", "query": "yearning", "prompt": "Process the full audio offline. Find every instance of the word \"yearning\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.15, 3.85, "yearning"]], "response_mode": "offline", "answer": "[{\"word\":\"yearning\",\"start\":3.15,\"end\":3.85}]"} {"utterance_id": "1995-1836-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0009.flac", "task_type": "lexical_detection", "query": "differs", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"differs\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.55, 4.24, "differs"]], "response_mode": "offline", "answer": "[{\"word\":\"differs\",\"start\":3.55,\"end\":4.24}]"} {"utterance_id": "5683-32879-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0010.flac", "task_type": "lexical_detection", "query": "events", "prompt": "When the utterance is complete, report all instances of the word \"events\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.96, 5.38, "events"]], "response_mode": "offline", "answer": "[{\"word\":\"events\",\"start\":4.96,\"end\":5.38}]"} {"utterance_id": "7729-102255-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0018.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.27, 2.56, "latter"]], "response_mode": "offline", "answer": "[{\"word\":\"latter\",\"start\":2.27,\"end\":2.56}]"} {"utterance_id": "2830-3980-0064", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0064.flac", "task_type": "lexical_detection", "query": "obtain", "prompt": "Once the audio has finished, identify every instance of the word \"obtain\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.06, 1.46, "obtain"]], "response_mode": "offline", "answer": "[{\"word\":\"obtain\",\"start\":1.06,\"end\":1.46}]"} {"utterance_id": "1580-141084-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0000.flac", "task_type": "semantic_detection", "query": "color", "prompt": "When the utterance is complete, report all instances of a color name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.81, 2.09, "dark"]], "response_mode": "offline", "answer": "[{\"word\":\"dark\",\"start\":1.81,\"end\":2.09}]"} {"utterance_id": "908-31957-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0016.flac", "task_type": "lexical_detection", "query": "dost", "prompt": "Once the audio has finished, identify every instance of the word \"dost\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.82, 5.26, "dost"]], "response_mode": "offline", "answer": "[{\"word\":\"dost\",\"start\":4.82,\"end\":5.26}]"} {"utterance_id": "2830-3980-0068", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0068.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen carefully to the entire clip, then report every instance of a family relation as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.43, 2.66, "son"]], "response_mode": "offline", "answer": "[{\"word\":\"son\",\"start\":2.43,\"end\":2.66}]"} {"utterance_id": "1284-1180-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0021.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "After the audio ends, extract all occurrences of an animal name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.49, 1.78, "cat"], [7.18, 7.56, "mice"]], "response_mode": "offline", "answer": "[{\"word\":\"cat\",\"start\":1.49,\"end\":1.78},{\"word\":\"mice\",\"start\":7.18,\"end\":7.56}]"} {"utterance_id": "1089-134691-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0016.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Once the audio has finished, identify every instance of a country name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.77, 7.27, "ireland"]], "response_mode": "offline", "answer": "[{\"word\":\"ireland\",\"start\":6.77,\"end\":7.27}]"} {"utterance_id": "8455-210777-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0008.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.18, 3.59, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":3.18,\"end\":3.59}]"} {"utterance_id": "4992-41806-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0007.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "When the utterance is complete, report how a person's name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.91, "one"], [5.8, 6.7, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.53,\"end\":0.91},{\"word\":\"two\",\"start\":5.8,\"end\":6.7}]"} {"utterance_id": "7021-85628-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0005.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.23, 4.66, "king"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":4.23,\"end\":4.66}]"} {"utterance_id": "4507-16021-0049", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0049.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a weather term with fields word, start, end (seconds). If none, reply [].", "labels": [[5.22, 5.68, "sunny"]], "response_mode": "offline", "answer": "[{\"word\":\"sunny\",\"start\":5.22,\"end\":5.68}]"} {"utterance_id": "1089-134686-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.5, 1.06, "stephen"], [6.38, 6.8, "heron"]], "response_mode": "offline", "answer": "[{\"word\":\"stephen\",\"start\":0.5,\"end\":1.06},{\"word\":\"heron\",\"start\":6.38,\"end\":6.8}]"} {"utterance_id": "2094-142345-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0000.flac", "task_type": "lexical_detection", "query": "brick", "prompt": "Once the audio has finished, identify every instance of the word \"brick\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.66, 3.06, "brick"], [12.81, 13.23, "brick"]], "response_mode": "offline", "answer": "[{\"word\":\"brick\",\"start\":2.66,\"end\":3.06},{\"word\":\"brick\",\"start\":12.81,\"end\":13.23}]"} {"utterance_id": "7127-75946-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0007.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[3.8, 4.06, "king"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":3.8,\"end\":4.06}]"} {"utterance_id": "237-134500-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0018.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen carefully to the entire clip, then report every instance of an animal name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.12, 9.64, "coral"]], "response_mode": "offline", "answer": "[{\"word\":\"coral\",\"start\":9.12,\"end\":9.64}]"} {"utterance_id": "1221-135767-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0018.flac", "task_type": "lexical_detection", "query": "mirror", "prompt": "After the audio ends, extract all occurrences of the word \"mirror\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.41, 9.85, "mirror"]], "response_mode": "offline", "answer": "[{\"word\":\"mirror\",\"start\":9.41,\"end\":9.85}]"} {"utterance_id": "7176-88083-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0013.flac", "task_type": "lexical_detection", "query": "smitten", "prompt": "After the audio ends, extract all occurrences of the word \"smitten\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.45, 4.91, "smitten"]], "response_mode": "offline", "answer": "[{\"word\":\"smitten\",\"start\":4.45,\"end\":4.91}]"} {"utterance_id": "3729-6852-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0042.flac", "task_type": "lexical_detection", "query": "asking", "prompt": "After listening to the whole recording, output every instance of the word \"asking\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.47, 3.94, "asking"]], "response_mode": "offline", "answer": "[{\"word\":\"asking\",\"start\":3.47,\"end\":3.94}]"} {"utterance_id": "4446-2273-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0000.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.62, 6.83, "chin"]], "response_mode": "offline", "answer": "[{\"word\":\"chin\",\"start\":6.62,\"end\":6.83}]"} {"utterance_id": "5142-36377-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0024.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[3.24, 3.55, "two"], [4.9, 5.1, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":3.24,\"end\":3.55},{\"word\":\"two\",\"start\":4.9,\"end\":5.1}]"} {"utterance_id": "3729-6852-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0041.flac", "task_type": "lexical_detection", "query": "principal", "prompt": "Once the audio has finished, identify every instance of the word \"principal\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.89, 5.41, "principal"]], "response_mode": "offline", "answer": "[{\"word\":\"principal\",\"start\":4.89,\"end\":5.41}]"} {"utterance_id": "2961-961-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0012.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "After the audio ends, extract all occurrences of a weather term. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.61, 3.12, "flood"]], "response_mode": "offline", "answer": "[{\"word\":\"flood\",\"start\":2.61,\"end\":3.12}]"} {"utterance_id": "2300-131720-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0018.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Process the full audio offline. Find every instance of a city name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.66, 12.09, "cities"]], "response_mode": "offline", "answer": "[{\"word\":\"cities\",\"start\":11.66,\"end\":12.09}]"} {"utterance_id": "4992-41797-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0015.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.47, 4.88, "animal"]], "response_mode": "offline", "answer": "[{\"word\":\"animal\",\"start\":4.47,\"end\":4.88}]"} {"utterance_id": "3575-170457-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0047.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.7, 4.12, "charlotte"]], "response_mode": "offline", "answer": "[{\"word\":\"charlotte\",\"start\":3.7,\"end\":4.12}]"} {"utterance_id": "1995-1837-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0016.flac", "task_type": "lexical_detection", "query": "amazement", "prompt": "When the utterance is complete, report all instances of the word \"amazement\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.8, 4.39, "amazement"]], "response_mode": "offline", "answer": "[{\"word\":\"amazement\",\"start\":3.8,\"end\":4.39}]"} {"utterance_id": "1995-1826-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0026.flac", "task_type": "lexical_detection", "query": "conversation", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"conversation\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[11.4, 12.22, "conversation"]], "response_mode": "offline", "answer": "[{\"word\":\"conversation\",\"start\":11.4,\"end\":12.22}]"} {"utterance_id": "5142-36377-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0013.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.54, 2.31, "john jago"]], "response_mode": "offline", "answer": "[{\"word\":\"john jago\",\"start\":1.54,\"end\":2.31}]"} {"utterance_id": "1995-1836-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0001.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an organization name with fields word, start, end (seconds). If none, reply [].", "labels": [[5.22, 5.72, "senate"]], "response_mode": "offline", "answer": "[{\"word\":\"senate\",\"start\":5.22,\"end\":5.72}]"} {"utterance_id": "1089-134686-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0029.flac", "task_type": "semantic_detection", "query": "date", "prompt": "When the utterance is complete, report all instances of a date including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 0.85, "friday"]], "response_mode": "offline", "answer": "[{\"word\":\"friday\",\"start\":0.42,\"end\":0.85}]"} {"utterance_id": "5142-33396-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0031.flac", "task_type": "lexical_detection", "query": "pot", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"pot\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.38, 2.63, "pot"]], "response_mode": "offline", "answer": "[{\"word\":\"pot\",\"start\":2.38,\"end\":2.63}]"} {"utterance_id": "4992-41806-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0014.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.88, 3.59, "osh popham"]], "response_mode": "offline", "answer": "[{\"word\":\"osh popham\",\"start\":2.88,\"end\":3.59}]"} {"utterance_id": "908-157963-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0022.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.81, 1.25, "worm"]], "response_mode": "offline", "answer": "[{\"word\":\"worm\",\"start\":0.81,\"end\":1.25}]"} {"utterance_id": "4446-2273-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0012.flac", "task_type": "lexical_detection", "query": "thank", "prompt": "Process the full audio offline. Find every instance of the word \"thank\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.35, 0.69, "thank"]], "response_mode": "offline", "answer": "[{\"word\":\"thank\",\"start\":0.35,\"end\":0.69}]"} {"utterance_id": "6829-68769-0049", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0049.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[6.68, 7.1, "gates"]], "response_mode": "offline", "answer": "[{\"word\":\"gates\",\"start\":6.68,\"end\":7.1}]"} {"utterance_id": "260-123286-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0003.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a family relation with fields word, start, end (seconds). If none, reply [].", "labels": [[1.63, 2.04, "uncle"]], "response_mode": "offline", "answer": "[{\"word\":\"uncle\",\"start\":1.63,\"end\":2.04}]"} {"utterance_id": "5683-32865-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0009.flac", "task_type": "lexical_detection", "query": "sly", "prompt": "After the audio ends, extract all occurrences of the word \"sly\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.21, 8.65, "sly"]], "response_mode": "offline", "answer": "[{\"word\":\"sly\",\"start\":8.21,\"end\":8.65}]"} {"utterance_id": "4970-29095-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0014.flac", "task_type": "lexical_detection", "query": "known", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"known\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.37, 2.89, "known"]], "response_mode": "offline", "answer": "[{\"word\":\"known\",\"start\":2.37,\"end\":2.89}]"} {"utterance_id": "5105-28240-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0020.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.55, 1.86, "head"]], "response_mode": "offline", "answer": "[{\"word\":\"head\",\"start\":1.55,\"end\":1.86}]"} {"utterance_id": "1580-141084-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0019.flac", "task_type": "lexical_detection", "query": "solved", "prompt": "When the utterance is complete, report all instances of the word \"solved\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.87, 2.23, "solved"]], "response_mode": "offline", "answer": "[{\"word\":\"solved\",\"start\":1.87,\"end\":2.23}]"} {"utterance_id": "3729-6852-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0019.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a food name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.57, 2.09, "coffee"]], "response_mode": "offline", "answer": "[{\"word\":\"coffee\",\"start\":1.57,\"end\":2.09}]"} {"utterance_id": "8463-294825-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0012.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen carefully to the entire clip, then report every instance of an animal name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.34, 0.94, "nautilus"]], "response_mode": "offline", "answer": "[{\"word\":\"nautilus\",\"start\":0.34,\"end\":0.94}]"} {"utterance_id": "8463-294828-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0015.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.93, "conseil"], [3.38, 3.95, "conseil"]], "response_mode": "offline", "answer": "[{\"word\":\"conseil\",\"start\":0.53,\"end\":0.93},{\"word\":\"conseil\",\"start\":3.38,\"end\":3.95}]"} {"utterance_id": "1188-133604-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0023.flac", "task_type": "lexical_detection", "query": "viper", "prompt": "Process the full audio offline. Find every instance of the word \"viper\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.96, 8.33, "viper"]], "response_mode": "offline", "answer": "[{\"word\":\"viper\",\"start\":7.96,\"end\":8.33}]"} {"utterance_id": "6829-68769-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0051.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.69, 3.19, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":2.69,\"end\":3.19}]"} {"utterance_id": "1320-122612-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0016.flac", "task_type": "lexical_detection", "query": "size", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"size\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.8, 2.21, "size"]], "response_mode": "offline", "answer": "[{\"word\":\"size\",\"start\":1.8,\"end\":2.21}]"} {"utterance_id": "1089-134691-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0003.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "When the utterance is complete, report all instances of an organization name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.72, 1.82, "university"]], "response_mode": "offline", "answer": "[{\"word\":\"university\",\"start\":0.72,\"end\":1.82}]"} {"utterance_id": "1580-141083-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0018.flac", "task_type": "lexical_detection", "query": "soames", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"soames\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.8, 1.32, "soames"]], "response_mode": "offline", "answer": "[{\"word\":\"soames\",\"start\":0.8,\"end\":1.32}]"} {"utterance_id": "8463-294828-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0032.flac", "task_type": "lexical_detection", "query": "aboard", "prompt": "After listening to the whole recording, output every instance of the word \"aboard\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.44, 1.8, "aboard"]], "response_mode": "offline", "answer": "[{\"word\":\"aboard\",\"start\":1.44,\"end\":1.8}]"} {"utterance_id": "7176-88083-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0004.flac", "task_type": "lexical_detection", "query": "torrent", "prompt": "When the utterance is complete, report all instances of the word \"torrent\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.06, 5.53, "torrent"]], "response_mode": "offline", "answer": "[{\"word\":\"torrent\",\"start\":5.06,\"end\":5.53}]"} {"utterance_id": "2094-142345-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0038.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.54, 12.93, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":12.54,\"end\":12.93}]"} {"utterance_id": "4970-29093-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0005.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After the audio ends, extract all occurrences of a plant name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.92, 15.55, "date palm"]], "response_mode": "offline", "answer": "[{\"word\":\"date palm\",\"start\":14.92,\"end\":15.55}]"} {"utterance_id": "1221-135766-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0013.flac", "task_type": "lexical_detection", "query": "infantile", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"infantile\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.2, 2.9, "infantile"]], "response_mode": "offline", "answer": "[{\"word\":\"infantile\",\"start\":2.2,\"end\":2.9}]"} {"utterance_id": "672-122797-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0042.flac", "task_type": "lexical_detection", "query": "loft", "prompt": "After listening to the whole recording, output every instance of the word \"loft\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.16, 6.73, "loft"]], "response_mode": "offline", "answer": "[{\"word\":\"loft\",\"start\":6.16,\"end\":6.73}]"} {"utterance_id": "2094-142345-0056", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0056.flac", "task_type": "lexical_detection", "query": "doing", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"doing\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.91, 1.18, "doing"]], "response_mode": "offline", "answer": "[{\"word\":\"doing\",\"start\":0.91,\"end\":1.18}]"} {"utterance_id": "8224-274384-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.76, 10.29, "david"]], "response_mode": "offline", "answer": "[{\"word\":\"david\",\"start\":9.76,\"end\":10.29}]"} {"utterance_id": "2830-3980-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0031.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "You will hear a complete utterance. When it ends, list every instance of a state or province name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.96, 2.54, "provinces"]], "response_mode": "offline", "answer": "[{\"word\":\"provinces\",\"start\":1.96,\"end\":2.54}]"} {"utterance_id": "1089-134686-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0017.flac", "task_type": "lexical_detection", "query": "stephen", "prompt": "Once the audio has finished, identify every instance of the word \"stephen\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.5, 1.06, "stephen"]], "response_mode": "offline", "answer": "[{\"word\":\"stephen\",\"start\":0.5,\"end\":1.06}]"} {"utterance_id": "61-70968-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0008.flac", "task_type": "lexical_detection", "query": "crystal", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"crystal\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.91, 2.33, "crystal"]], "response_mode": "offline", "answer": "[{\"word\":\"crystal\",\"start\":1.91,\"end\":2.33}]"} {"utterance_id": "1995-1826-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0025.flac", "task_type": "lexical_detection", "query": "you'll", "prompt": "After the audio ends, extract all occurrences of the word \"you'll\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.1, 1.22, "you'll"]], "response_mode": "offline", "answer": "[{\"word\":\"you'll\",\"start\":1.1,\"end\":1.22}]"} {"utterance_id": "3570-5694-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0017.flac", "task_type": "lexical_detection", "query": "uniforms", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"uniforms\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.89, 1.46, "uniforms"]], "response_mode": "offline", "answer": "[{\"word\":\"uniforms\",\"start\":0.89,\"end\":1.46}]"} {"utterance_id": "8224-274381-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0017.flac", "task_type": "lexical_detection", "query": "hurrying", "prompt": "After listening to the whole recording, output every instance of the word \"hurrying\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.96, 12.55, "hurrying"]], "response_mode": "offline", "answer": "[{\"word\":\"hurrying\",\"start\":11.96,\"end\":12.55}]"} {"utterance_id": "7729-102255-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0028.flac", "task_type": "lexical_detection", "query": "destruction", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"destruction\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.99, 7.86, "destruction"]], "response_mode": "offline", "answer": "[{\"word\":\"destruction\",\"start\":6.99,\"end\":7.86}]"} {"utterance_id": "7729-102255-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0015.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "Listen to the full audio. Count every occurrence of a profession. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.36, 2.01, "one"], [3.24, 3.87, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.36,\"end\":2.01},{\"word\":\"two\",\"start\":3.24,\"end\":3.87}]"} {"utterance_id": "5639-40744-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0004.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.1, 2.7, "faces"]], "response_mode": "offline", "answer": "[{\"word\":\"faces\",\"start\":2.1,\"end\":2.7}]"} {"utterance_id": "2961-960-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0004.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.37, 3.32, "commentators"]], "response_mode": "offline", "answer": "[{\"word\":\"commentators\",\"start\":2.37,\"end\":3.32}]"} {"utterance_id": "7729-102255-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0016.flac", "task_type": "semantic_counting", "query": "organization", "prompt": "Process the full audio offline. For every occurrence of an organization name, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.24, 5.73, "one"], [10.87, 11.46, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.24,\"end\":5.73},{\"word\":\"two\",\"start\":10.87,\"end\":11.46}]"} {"utterance_id": "7127-75947-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0039.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.93, 9.66, "cavaliers"]], "response_mode": "offline", "answer": "[{\"word\":\"cavaliers\",\"start\":8.93,\"end\":9.66}]"} {"utterance_id": "1089-134686-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0019.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[4.62, 4.89, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.62,\"end\":4.89}]"} {"utterance_id": "4970-29095-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0026.flac", "task_type": "lexical_detection", "query": "quiet", "prompt": "After listening to the whole recording, output every instance of the word \"quiet\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.61, 4.09, "quiet"]], "response_mode": "offline", "answer": "[{\"word\":\"quiet\",\"start\":3.61,\"end\":4.09}]"} {"utterance_id": "1995-1826-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0016.flac", "task_type": "lexical_detection", "query": "delicate", "prompt": "Once the audio has finished, identify every instance of the word \"delicate\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.32, 1.76, "delicate"]], "response_mode": "offline", "answer": "[{\"word\":\"delicate\",\"start\":1.32,\"end\":1.76}]"} {"utterance_id": "6829-68769-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0008.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Listen to the full audio. After it finishes, find every instance of a state or province name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.57, 3.9, "county"]], "response_mode": "offline", "answer": "[{\"word\":\"county\",\"start\":3.57,\"end\":3.9}]"} {"utterance_id": "1995-1826-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0002.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Do not answer until the audio ends. Then list each instance of an organization name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.17, 2.64, "college"]], "response_mode": "offline", "answer": "[{\"word\":\"college\",\"start\":2.17,\"end\":2.64}]"} {"utterance_id": "8463-294828-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0009.flac", "task_type": "lexical_detection", "query": "once", "prompt": "Once the audio has finished, identify every instance of the word \"once\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.67, 0.98, "once"]], "response_mode": "offline", "answer": "[{\"word\":\"once\",\"start\":0.67,\"end\":0.98}]"} {"utterance_id": "4446-2273-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0034.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the full audio. After it finishes, find every instance of a color name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.94, 2.17, "yellow"]], "response_mode": "offline", "answer": "[{\"word\":\"yellow\",\"start\":1.94,\"end\":2.17}]"} {"utterance_id": "5683-32879-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0020.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.4, 4.9, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":4.4,\"end\":4.9}]"} {"utterance_id": "3729-6852-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0034.flac", "task_type": "lexical_detection", "query": "birth", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"birth\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[15.8, 16.29, "birth"]], "response_mode": "offline", "answer": "[{\"word\":\"birth\",\"start\":15.8,\"end\":16.29}]"} {"utterance_id": "8455-210777-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0007.flac", "task_type": "lexical_detection", "query": "eva", "prompt": "When the utterance is complete, report all instances of the word \"eva\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.8, 2.23, "eva"]], "response_mode": "offline", "answer": "[{\"word\":\"eva\",\"start\":1.8,\"end\":2.23}]"} {"utterance_id": "4992-23283-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0010.flac", "task_type": "lexical_detection", "query": "election", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"election\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.17, 2.58, "election"]], "response_mode": "offline", "answer": "[{\"word\":\"election\",\"start\":2.17,\"end\":2.58}]"} {"utterance_id": "2830-3980-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0029.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.99, 2.5, "apostles"]], "response_mode": "offline", "answer": "[{\"word\":\"apostles\",\"start\":1.99,\"end\":2.5}]"} {"utterance_id": "908-157963-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0016.flac", "task_type": "lexical_detection", "query": "hears", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"hears\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.53, 3.94, "hears"]], "response_mode": "offline", "answer": "[{\"word\":\"hears\",\"start\":3.53,\"end\":3.94}]"} {"utterance_id": "6930-76324-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0010.flac", "task_type": "lexical_detection", "query": "world", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"world\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.81, 1.11, "world"]], "response_mode": "offline", "answer": "[{\"word\":\"world\",\"start\":0.81,\"end\":1.11}]"} {"utterance_id": "6829-68771-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0029.flac", "task_type": "lexical_detection", "query": "aware", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"aware\" with fields word, start, end (seconds). If none, reply [].", "labels": [[7.78, 8.09, "aware"]], "response_mode": "offline", "answer": "[{\"word\":\"aware\",\"start\":7.78,\"end\":8.09}]"} {"utterance_id": "1221-135767-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0024.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 0.9, "pearl"]], "response_mode": "offline", "answer": "[{\"word\":\"pearl\",\"start\":0.45,\"end\":0.9}]"} {"utterance_id": "3729-6852-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0000.flac", "task_type": "lexical_detection", "query": "arrival", "prompt": "Once the audio has finished, identify every instance of the word \"arrival\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.32, 1.81, "arrival"]], "response_mode": "offline", "answer": "[{\"word\":\"arrival\",\"start\":1.32,\"end\":1.81}]"} {"utterance_id": "121-123852-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123852/121-123852-0000.flac", "task_type": "lexical_detection", "query": "sometime", "prompt": "Once the audio has finished, identify every instance of the word \"sometime\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.16, 6.03, "sometime"]], "response_mode": "offline", "answer": "[{\"word\":\"sometime\",\"start\":5.16,\"end\":6.03}]"} {"utterance_id": "61-70968-0061", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0061.flac", "task_type": "semantic_detection", "query": "food", "prompt": "You will hear a complete utterance. When it ends, list every instance of a food name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.92, 4.23, "wine"]], "response_mode": "offline", "answer": "[{\"word\":\"wine\",\"start\":3.92,\"end\":4.23}]"} {"utterance_id": "5639-40744-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0039.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.29, 4.81, "arms"], [11.77, 12.22, "arms"]], "response_mode": "offline", "answer": "[{\"word\":\"arms\",\"start\":4.29,\"end\":4.81},{\"word\":\"arms\",\"start\":11.77,\"end\":12.22}]"} {"utterance_id": "1221-135767-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0003.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "After listening to the whole recording, output every instance of a state or province name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.05, 12.61, "colony"]], "response_mode": "offline", "answer": "[{\"word\":\"colony\",\"start\":12.05,\"end\":12.61}]"} {"utterance_id": "3575-170457-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0013.flac", "task_type": "lexical_detection", "query": "proper", "prompt": "Process the full audio offline. Find every instance of the word \"proper\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.8, 2.16, "proper"]], "response_mode": "offline", "answer": "[{\"word\":\"proper\",\"start\":1.8,\"end\":2.16}]"} {"utterance_id": "8455-210777-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0014.flac", "task_type": "lexical_detection", "query": "player", "prompt": "After the audio ends, extract all occurrences of the word \"player\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.43, 2.84, "player"]], "response_mode": "offline", "answer": "[{\"word\":\"player\",\"start\":2.43,\"end\":2.84}]"} {"utterance_id": "1221-135766-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0005.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 0.98, "hester"], [10.32, 10.69, "pearl"]], "response_mode": "offline", "answer": "[{\"word\":\"hester\",\"start\":0.51,\"end\":0.98},{\"word\":\"pearl\",\"start\":10.32,\"end\":10.69}]"} {"utterance_id": "3729-6852-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0018.flac", "task_type": "lexical_detection", "query": "wishes", "prompt": "When the utterance is complete, report all instances of the word \"wishes\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.4, 5.93, "wishes"]], "response_mode": "offline", "answer": "[{\"word\":\"wishes\",\"start\":5.4,\"end\":5.93}]"} {"utterance_id": "237-134500-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0000.flac", "task_type": "lexical_counting", "query": "read", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"read\" (word, start, end). If none, reply [].", "labels": [[0.71, 0.94, "one"], [2.83, 3.05, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.71,\"end\":0.94},{\"word\":\"two\",\"start\":2.83,\"end\":3.05}]"} {"utterance_id": "8555-292519-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0003.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Do not answer until the audio ends. Then list each instance of a time of day as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.76, 1.48, "sunset"]], "response_mode": "offline", "answer": "[{\"word\":\"sunset\",\"start\":0.76,\"end\":1.48}]"} {"utterance_id": "2830-3979-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0001.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Process the full audio offline. Find every instance of an organization name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.47, 10.39, "lutheran church"]], "response_mode": "offline", "answer": "[{\"word\":\"lutheran church\",\"start\":9.47,\"end\":10.39}]"} {"utterance_id": "8463-294828-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0006.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.02, 1.48, "shoulders"]], "response_mode": "offline", "answer": "[{\"word\":\"shoulders\",\"start\":1.02,\"end\":1.48}]"} {"utterance_id": "4077-13751-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0003.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Listen to the full audio. Count every occurrence of a number. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.03, 1.34, "one"], [7.27, 7.7, "two"], [9.29, 10.02, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.03,\"end\":1.34},{\"word\":\"two\",\"start\":7.27,\"end\":7.7},{\"word\":\"three\",\"start\":9.29,\"end\":10.02}]"} {"utterance_id": "7176-92135-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0008.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.72, 0.93, "ear"]], "response_mode": "offline", "answer": "[{\"word\":\"ear\",\"start\":0.72,\"end\":0.93}]"} {"utterance_id": "121-121726-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0005.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Do not answer until the audio ends. Then list each instance of a plant name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.52, 1.09, "hedge"]], "response_mode": "offline", "answer": "[{\"word\":\"hedge\",\"start\":0.52,\"end\":1.09}]"} {"utterance_id": "1284-1180-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0004.flac", "task_type": "lexical_detection", "query": "path", "prompt": "Process the full audio offline. Find every instance of the word \"path\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.6, 4.11, "path"]], "response_mode": "offline", "answer": "[{\"word\":\"path\",\"start\":3.6,\"end\":4.11}]"} {"utterance_id": "5105-28241-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0006.flac", "task_type": "lexical_detection", "query": "log", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"log\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.52, 0.86, "log"]], "response_mode": "offline", "answer": "[{\"word\":\"log\",\"start\":0.52,\"end\":0.86}]"} {"utterance_id": "3729-6852-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0045.flac", "task_type": "lexical_counting", "query": "good", "prompt": "Listen to the entire clip, then report the count sequence for the word \"good\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.88, 1.09, "one"], [2.4, 2.63, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.88,\"end\":1.09},{\"word\":\"two\",\"start\":2.4,\"end\":2.63}]"} {"utterance_id": "3575-170457-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0015.flac", "task_type": "lexical_detection", "query": "discourage", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"discourage\" with fields word, start, end (seconds). If none, reply [].", "labels": [[5.66, 6.16, "discourage"]], "response_mode": "offline", "answer": "[{\"word\":\"discourage\",\"start\":5.66,\"end\":6.16}]"} {"utterance_id": "8455-210777-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0037.flac", "task_type": "lexical_detection", "query": "threatening", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"threatening\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.33, 1.78, "threatening"]], "response_mode": "offline", "answer": "[{\"word\":\"threatening\",\"start\":1.33,\"end\":1.78}]"} {"utterance_id": "4507-16021-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0053.flac", "task_type": "lexical_counting", "query": "number", "prompt": "Process the full audio offline. For every occurrence of the word \"number\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.41, 1.91, "one"], [4.54, 5.03, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.41,\"end\":1.91},{\"word\":\"two\",\"start\":4.54,\"end\":5.03}]"} {"utterance_id": "3570-5695-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0004.flac", "task_type": "lexical_counting", "query": "achieved", "prompt": "After listening to the whole recording, output timed count words for the word \"achieved\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[1.3, 1.81, "one"], [5.57, 6.0, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.3,\"end\":1.81},{\"word\":\"two\",\"start\":5.57,\"end\":6.0}]"} {"utterance_id": "260-123440-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0016.flac", "task_type": "lexical_detection", "query": "drowned", "prompt": "When the utterance is complete, report all instances of the word \"drowned\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.28, 3.72, "drowned"]], "response_mode": "offline", "answer": "[{\"word\":\"drowned\",\"start\":3.28,\"end\":3.72}]"} {"utterance_id": "2961-960-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0020.flac", "task_type": "lexical_detection", "query": "hence", "prompt": "After listening to the whole recording, output every instance of the word \"hence\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.61, 1.21, "hence"]], "response_mode": "offline", "answer": "[{\"word\":\"hence\",\"start\":0.61,\"end\":1.21}]"} {"utterance_id": "7021-85628-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0006.flac", "task_type": "lexical_detection", "query": "ball", "prompt": "When the utterance is complete, report all instances of the word \"ball\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.58, 2.11, "ball"]], "response_mode": "offline", "answer": "[{\"word\":\"ball\",\"start\":1.58,\"end\":2.11}]"} {"utterance_id": "4992-41797-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0018.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.51, 4.38, "olive lord"]], "response_mode": "offline", "answer": "[{\"word\":\"olive lord\",\"start\":3.51,\"end\":4.38}]"} {"utterance_id": "1089-134686-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0022.flac", "task_type": "semantic_counting", "query": "ordinal", "prompt": "Listen to the audio in full, then emit a JSON array of count events for an ordinal number (word, start, end). If none, reply [].", "labels": [[1.79, 2.21, "one"], [5.95, 6.47, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.79,\"end\":2.21},{\"word\":\"two\",\"start\":5.95,\"end\":6.47}]"} {"utterance_id": "8455-210777-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0030.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "After listening to the whole recording, output timed count words for a profession (one, two, three, ...) as JSON. If none, reply [].", "labels": [[2.44, 3.04, "one"], [7.26, 7.79, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.44,\"end\":3.04},{\"word\":\"two\",\"start\":7.26,\"end\":7.79}]"} {"utterance_id": "4970-29093-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0018.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.51, 6.88, "uncle"]], "response_mode": "offline", "answer": "[{\"word\":\"uncle\",\"start\":6.51,\"end\":6.88}]"} {"utterance_id": "61-70968-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0014.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Once the audio has finished, identify every instance of a color name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.42, 5.77, "silver"]], "response_mode": "offline", "answer": "[{\"word\":\"silver\",\"start\":5.42,\"end\":5.77}]"} {"utterance_id": "8463-294828-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0019.flac", "task_type": "lexical_detection", "query": "whole", "prompt": "After listening to the whole recording, output every instance of the word \"whole\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.64, 2.9, "whole"]], "response_mode": "offline", "answer": "[{\"word\":\"whole\",\"start\":2.64,\"end\":2.9}]"} {"utterance_id": "5639-40744-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0031.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.64, 10.89, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":10.64,\"end\":10.89}]"} {"utterance_id": "61-70970-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0012.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "You will hear a complete utterance. When it ends, list every instance of a time of day you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.96, 2.34, "morning"]], "response_mode": "offline", "answer": "[{\"word\":\"morning\",\"start\":1.96,\"end\":2.34}]"} {"utterance_id": "7176-88083-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0010.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Do not answer until the audio ends. Then list each instance of an animal name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.33, 0.7, "cat"]], "response_mode": "offline", "answer": "[{\"word\":\"cat\",\"start\":0.33,\"end\":0.7}]"} {"utterance_id": "121-127105-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0004.flac", "task_type": "lexical_detection", "query": "written", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"written\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.3, 1.85, "written"]], "response_mode": "offline", "answer": "[{\"word\":\"written\",\"start\":1.3,\"end\":1.85}]"} {"utterance_id": "7127-75947-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0040.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Once the audio has finished, identify every instance of an animal name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[7.79, 8.03, "deer"], [9.21, 9.61, "pheasant"], [10.9, 11.25, "fox"]], "response_mode": "offline", "answer": "[{\"word\":\"deer\",\"start\":7.79,\"end\":8.03},{\"word\":\"pheasant\",\"start\":9.21,\"end\":9.61},{\"word\":\"fox\",\"start\":10.9,\"end\":11.25}]"} {"utterance_id": "8455-210777-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0041.flac", "task_type": "lexical_detection", "query": "shock", "prompt": "After listening to the whole recording, output every instance of the word \"shock\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.77, 2.18, "shock"]], "response_mode": "offline", "answer": "[{\"word\":\"shock\",\"start\":1.77,\"end\":2.18}]"} {"utterance_id": "1320-122617-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0039.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.58, 1.97, "feet"]], "response_mode": "offline", "answer": "[{\"word\":\"feet\",\"start\":1.58,\"end\":1.97}]"} {"utterance_id": "237-134493-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0007.flac", "task_type": "lexical_detection", "query": "alexandra", "prompt": "When the utterance is complete, report all instances of the word \"alexandra\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.18, 0.83, "alexandra"]], "response_mode": "offline", "answer": "[{\"word\":\"alexandra\",\"start\":0.18,\"end\":0.83}]"} {"utterance_id": "2094-142345-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0002.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[9.66, 10.26, "lackey"]], "response_mode": "offline", "answer": "[{\"word\":\"lackey\",\"start\":9.66,\"end\":10.26}]"} {"utterance_id": "2961-960-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0000.flac", "task_type": "lexical_detection", "query": "dreaming", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"dreaming\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[21.79, 22.41, "dreaming"]], "response_mode": "offline", "answer": "[{\"word\":\"dreaming\",\"start\":21.79,\"end\":22.41}]"} {"utterance_id": "5639-40744-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0030.flac", "task_type": "lexical_counting", "query": "leocadia", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"leocadia\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.96, 1.56, "one"], [13.09, 13.65, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.96,\"end\":1.56},{\"word\":\"two\",\"start\":13.09,\"end\":13.65}]"} {"utterance_id": "4077-13754-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0006.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, list every instance of an organization name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.73, 1.42, "latter day saints"], [4.05, 4.55, "church"]], "response_mode": "offline", "answer": "[{\"word\":\"latter day saints\",\"start\":0.73,\"end\":1.42},{\"word\":\"church\",\"start\":4.05,\"end\":4.55}]"} {"utterance_id": "2300-131720-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0003.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[15.05, 15.69, "seventeen"]], "response_mode": "offline", "answer": "[{\"word\":\"seventeen\",\"start\":15.05,\"end\":15.69}]"} {"utterance_id": "61-70968-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0027.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.97, 1.39, "tumbler"], [2.63, 3.03, "squire"]], "response_mode": "offline", "answer": "[{\"word\":\"tumbler\",\"start\":0.97,\"end\":1.39},{\"word\":\"squire\",\"start\":2.63,\"end\":3.03}]"} {"utterance_id": "121-127105-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0021.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.31, 1.88, "douglas"]], "response_mode": "offline", "answer": "[{\"word\":\"douglas\",\"start\":1.31,\"end\":1.88}]"} {"utterance_id": "1188-133604-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0014.flac", "task_type": "lexical_detection", "query": "gothic", "prompt": "When the utterance is complete, report all instances of the word \"gothic\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.41, 2.81, "gothic"]], "response_mode": "offline", "answer": "[{\"word\":\"gothic\",\"start\":2.41,\"end\":2.81}]"} {"utterance_id": "7176-88083-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0014.flac", "task_type": "lexical_detection", "query": "quarry", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"quarry\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.63, 3.11, "quarry"]], "response_mode": "offline", "answer": "[{\"word\":\"quarry\",\"start\":2.63,\"end\":3.11}]"} {"utterance_id": "61-70968-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0025.flac", "task_type": "lexical_detection", "query": "voice", "prompt": "After the audio ends, extract all occurrences of the word \"voice\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.38, 3.63, "voice"]], "response_mode": "offline", "answer": "[{\"word\":\"voice\",\"start\":3.38,\"end\":3.63}]"} {"utterance_id": "1320-122617-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0024.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.34, 0.84, "hawkeye"]], "response_mode": "offline", "answer": "[{\"word\":\"hawkeye\",\"start\":0.34,\"end\":0.84}]"} {"utterance_id": "5683-32865-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0012.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "When the utterance is complete, report how a person's name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.46, 0.89, "one"], [10.42, 10.82, "two"], [18.68, 19.55, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.46,\"end\":0.89},{\"word\":\"two\",\"start\":10.42,\"end\":10.82},{\"word\":\"three\",\"start\":18.68,\"end\":19.55}]"} {"utterance_id": "8224-274381-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0002.flac", "task_type": "semantic_counting", "query": "organization", "prompt": "After the audio ends, tally every instance of an organization name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[2.7, 4.25, "one"], [7.25, 8.3, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.7,\"end\":4.25},{\"word\":\"two\",\"start\":7.25,\"end\":8.3}]"} {"utterance_id": "7021-79740-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0001.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a person's name with word/start/end. If none, reply [].", "labels": [[0.52, 0.92, "one"], [2.66, 3.13, "two"], [5.26, 5.79, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.52,\"end\":0.92},{\"word\":\"two\",\"start\":2.66,\"end\":3.13},{\"word\":\"three\",\"start\":5.26,\"end\":5.79}]"} {"utterance_id": "6829-68771-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 0.9, "louise"]], "response_mode": "offline", "answer": "[{\"word\":\"louise\",\"start\":0.45,\"end\":0.9}]"} {"utterance_id": "121-127105-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0016.flac", "task_type": "lexical_detection", "query": "post", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"post\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.44, 1.95, "post"]], "response_mode": "offline", "answer": "[{\"word\":\"post\",\"start\":1.44,\"end\":1.95}]"} {"utterance_id": "7021-85628-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0010.flac", "task_type": "lexical_detection", "query": "cakes", "prompt": "Once the audio has finished, identify every instance of the word \"cakes\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.86, 4.41, "cakes"]], "response_mode": "offline", "answer": "[{\"word\":\"cakes\",\"start\":3.86,\"end\":4.41}]"} {"utterance_id": "237-134500-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0013.flac", "task_type": "lexical_detection", "query": "picked", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"picked\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.78, 1.1, "picked"]], "response_mode": "offline", "answer": "[{\"word\":\"picked\",\"start\":0.78,\"end\":1.1}]"} {"utterance_id": "6829-68769-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.62, 1.03, "tom"]], "response_mode": "offline", "answer": "[{\"word\":\"tom\",\"start\":0.62,\"end\":1.03}]"} {"utterance_id": "2961-961-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0008.flac", "task_type": "lexical_detection", "query": "owing", "prompt": "Process the full audio offline. Find every instance of the word \"owing\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.16, 3.5, "owing"]], "response_mode": "offline", "answer": "[{\"word\":\"owing\",\"start\":3.16,\"end\":3.5}]"} {"utterance_id": "672-122797-0069", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0069.flac", "task_type": "lexical_detection", "query": "weeds", "prompt": "Process the full audio offline. Find every instance of the word \"weeds\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.24, 3.73, "weeds"]], "response_mode": "offline", "answer": "[{\"word\":\"weeds\",\"start\":3.24,\"end\":3.73}]"} {"utterance_id": "4992-41806-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0005.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.26, 1.76, "olive's"]], "response_mode": "offline", "answer": "[{\"word\":\"olive's\",\"start\":1.26,\"end\":1.76}]"} {"utterance_id": "4992-41797-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0016.flac", "task_type": "lexical_detection", "query": "they're", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"they're\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.44, 1.57, "they're"]], "response_mode": "offline", "answer": "[{\"word\":\"they're\",\"start\":1.44,\"end\":1.57}]"} {"utterance_id": "3575-170457-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0030.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.3, 6.81, "poet"]], "response_mode": "offline", "answer": "[{\"word\":\"poet\",\"start\":6.3,\"end\":6.81}]"} {"utterance_id": "7021-85628-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0025.flac", "task_type": "lexical_detection", "query": "close", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"close\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.82, 2.47, "close"]], "response_mode": "offline", "answer": "[{\"word\":\"close\",\"start\":1.82,\"end\":2.47}]"} {"utterance_id": "7127-75946-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0015.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.66, 4.47, "spring"]], "response_mode": "offline", "answer": "[{\"word\":\"spring\",\"start\":3.66,\"end\":4.47}]"} {"utterance_id": "6829-68769-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0021.flac", "task_type": "lexical_detection", "query": "gray", "prompt": "After listening to the whole recording, output every instance of the word \"gray\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.47, 3.77, "gray"]], "response_mode": "offline", "answer": "[{\"word\":\"gray\",\"start\":3.47,\"end\":3.77}]"} {"utterance_id": "1188-133604-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0018.flac", "task_type": "lexical_counting", "query": "art", "prompt": "Once the audio has finished, list each occurrence of the word \"art\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[1.67, 1.91, "one"], [10.74, 11.15, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.67,\"end\":1.91},{\"word\":\"two\",\"start\":10.74,\"end\":11.15}]"} {"utterance_id": "8230-279154-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0024.flac", "task_type": "lexical_detection", "query": "past", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"past\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.52, 6.16, "past"]], "response_mode": "offline", "answer": "[{\"word\":\"past\",\"start\":5.52,\"end\":6.16}]"} {"utterance_id": "5105-28241-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0015.flac", "task_type": "semantic_detection", "query": "city", "prompt": "You will hear a complete utterance. When it ends, list every instance of a city name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[25.18, 25.61, "city"]], "response_mode": "offline", "answer": "[{\"word\":\"city\",\"start\":25.18,\"end\":25.61}]"} {"utterance_id": "1995-1836-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0012.flac", "task_type": "lexical_detection", "query": "we're", "prompt": "After listening to the whole recording, output every instance of the word \"we're\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.85, 1.13, "we're"]], "response_mode": "offline", "answer": "[{\"word\":\"we're\",\"start\":0.85,\"end\":1.13}]"} {"utterance_id": "3575-170457-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0037.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.36, 6.67, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":6.36,\"end\":6.67}]"} {"utterance_id": "3729-6852-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0034.flac", "task_type": "lexical_counting", "query": "mad", "prompt": "Listen to the entire clip, then report the count sequence for the word \"mad\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.38, 7.86, "one"], [14.91, 15.32, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.38,\"end\":7.86},{\"word\":\"two\",\"start\":14.91,\"end\":15.32}]"} {"utterance_id": "8224-274381-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0004.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[18.31, 18.79, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":18.31,\"end\":18.79}]"} {"utterance_id": "2300-131720-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0011.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Once the audio has finished, identify every instance of a city name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.12, 3.75, "suburb"]], "response_mode": "offline", "answer": "[{\"word\":\"suburb\",\"start\":3.12,\"end\":3.75}]"} {"utterance_id": "8463-294825-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0008.flac", "task_type": "lexical_detection", "query": "nemo", "prompt": "Process the full audio offline. Find every instance of the word \"nemo\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.31, 3.82, "nemo"]], "response_mode": "offline", "answer": "[{\"word\":\"nemo\",\"start\":3.31,\"end\":3.82}]"} {"utterance_id": "121-127105-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0008.flac", "task_type": "lexical_detection", "query": "woman's", "prompt": "After listening to the whole recording, output every instance of the word \"woman's\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.04, 2.62, "woman's"]], "response_mode": "offline", "answer": "[{\"word\":\"woman's\",\"start\":2.04,\"end\":2.62}]"} {"utterance_id": "2830-3980-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0050.flac", "task_type": "lexical_detection", "query": "unto", "prompt": "Process the full audio offline. Find every instance of the word \"unto\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.72, 5.92, "unto"]], "response_mode": "offline", "answer": "[{\"word\":\"unto\",\"start\":5.72,\"end\":5.92}]"} {"utterance_id": "3729-6852-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0040.flac", "task_type": "lexical_detection", "query": "progress", "prompt": "After the audio ends, extract all occurrences of the word \"progress\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.42, 10.26, "progress"]], "response_mode": "offline", "answer": "[{\"word\":\"progress\",\"start\":9.42,\"end\":10.26}]"} {"utterance_id": "4970-29095-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0001.flac", "task_type": "lexical_detection", "query": "song", "prompt": "After listening to the whole recording, output every instance of the word \"song\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.99, 5.52, "song"]], "response_mode": "offline", "answer": "[{\"word\":\"song\",\"start\":4.99,\"end\":5.52}]"} {"utterance_id": "4446-2273-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0014.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 0.73, "few"]], "response_mode": "offline", "answer": "[{\"word\":\"few\",\"start\":0.52,\"end\":0.73}]"} {"utterance_id": "5639-40744-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0039.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a body part (word, start, end). If none, reply [].", "labels": [[4.29, 4.81, "one"], [11.77, 12.22, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.29,\"end\":4.81},{\"word\":\"two\",\"start\":11.77,\"end\":12.22}]"} {"utterance_id": "1284-134647-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0006.flac", "task_type": "lexical_detection", "query": "disgrace", "prompt": "After listening to the whole recording, output every instance of the word \"disgrace\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.56, 5.26, "disgrace"]], "response_mode": "offline", "answer": "[{\"word\":\"disgrace\",\"start\":4.56,\"end\":5.26}]"} {"utterance_id": "1580-141084-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0033.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.62, 1.96, "holmes"]], "response_mode": "offline", "answer": "[{\"word\":\"holmes\",\"start\":1.62,\"end\":1.96}]"} {"utterance_id": "5105-28233-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28233/5105-28233-0004.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the full audio. After it finishes, find every instance of an organization name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.18, 2.8, "detachment"]], "response_mode": "offline", "answer": "[{\"word\":\"detachment\",\"start\":2.18,\"end\":2.8}]"} {"utterance_id": "7021-85628-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0011.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Do not answer until the audio ends. Then list each instance of a plant name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.04, 4.66, "roses"]], "response_mode": "offline", "answer": "[{\"word\":\"roses\",\"start\":4.04,\"end\":4.66}]"} {"utterance_id": "3575-170457-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0036.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "After listening to the whole recording, output timed count words for a body part (one, two, three, ...) as JSON. If none, reply [].", "labels": [[0.73, 1.05, "one"], [19.66, 20.22, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.73,\"end\":1.05},{\"word\":\"two\",\"start\":19.66,\"end\":20.22}]"} {"utterance_id": "1188-133604-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0007.flac", "task_type": "lexical_counting", "query": "touch", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"touch\" with word/start/end. If none, reply [].", "labels": [[3.98, 4.23, "one"], [10.76, 11.06, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.98,\"end\":4.23},{\"word\":\"two\",\"start\":10.76,\"end\":11.06}]"} {"utterance_id": "1221-135766-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.66, 2.46, "hester prynne"]], "response_mode": "offline", "answer": "[{\"word\":\"hester prynne\",\"start\":1.66,\"end\":2.46}]"} {"utterance_id": "5639-40744-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0022.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[11.53, 11.78, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":11.53,\"end\":11.78}]"} {"utterance_id": "5683-32865-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0008.flac", "task_type": "lexical_detection", "query": "jewels", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"jewels\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.09, 5.76, "jewels"]], "response_mode": "offline", "answer": "[{\"word\":\"jewels\",\"start\":5.09,\"end\":5.76}]"} {"utterance_id": "1320-122617-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0037.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Once the audio has finished, identify every instance of a state or province name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.59, 1.04, "delaware"]], "response_mode": "offline", "answer": "[{\"word\":\"delaware\",\"start\":0.59,\"end\":1.04}]"} {"utterance_id": "8463-294825-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0019.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.41, 2.01, "one twenty"]], "response_mode": "offline", "answer": "[{\"word\":\"one twenty\",\"start\":1.41,\"end\":2.01}]"} {"utterance_id": "4992-41806-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0011.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "After listening to the whole recording, output timed count words for a person's name (one, two, three, ...) as JSON. If none, reply [].", "labels": [[0.35, 1.1, "one"], [2.29, 2.7, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.35,\"end\":1.1},{\"word\":\"two\",\"start\":2.29,\"end\":2.7}]"} {"utterance_id": "4446-2271-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0002.flac", "task_type": "lexical_detection", "query": "it's", "prompt": "When the utterance is complete, report all instances of the word \"it's\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.26, 0.47, "it's"]], "response_mode": "offline", "answer": "[{\"word\":\"it's\",\"start\":0.26,\"end\":0.47}]"} {"utterance_id": "4446-2273-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0003.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a person's name: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.71, 1.05, "one"], [3.33, 3.84, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.71,\"end\":1.05},{\"word\":\"two\",\"start\":3.33,\"end\":3.84}]"} {"utterance_id": "2094-142345-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0008.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Once the audio has finished, identify every instance of a plant name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[15.08, 15.69, "moss"]], "response_mode": "offline", "answer": "[{\"word\":\"moss\",\"start\":15.08,\"end\":15.69}]"} {"utterance_id": "3570-5694-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0006.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the full audio. After it finishes, find every instance of an ordinal number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.13, 9.49, "second"]], "response_mode": "offline", "answer": "[{\"word\":\"second\",\"start\":9.13,\"end\":9.49}]"} {"utterance_id": "7127-75947-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0038.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.81, 2.29, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":1.81,\"end\":2.29}]"} {"utterance_id": "5142-33396-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0003.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.21, 1.75, "viking"]], "response_mode": "offline", "answer": "[{\"word\":\"viking\",\"start\":1.21,\"end\":1.75}]"} {"utterance_id": "8555-292519-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0008.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Once the audio has finished, identify every instance of a city name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.67, 2.04, "city"]], "response_mode": "offline", "answer": "[{\"word\":\"city\",\"start\":1.67,\"end\":2.04}]"} {"utterance_id": "1580-141083-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0030.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.35, 1.04, "mister soames"]], "response_mode": "offline", "answer": "[{\"word\":\"mister soames\",\"start\":0.35,\"end\":1.04}]"} {"utterance_id": "4992-41806-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0007.flac", "task_type": "lexical_detection", "query": "fitted", "prompt": "After listening to the whole recording, output every instance of the word \"fitted\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.88, 9.28, "fitted"]], "response_mode": "offline", "answer": "[{\"word\":\"fitted\",\"start\":8.88,\"end\":9.28}]"} {"utterance_id": "672-122797-0066", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0066.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen carefully to the entire clip, then report every instance of a time of day as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.99, 1.3, "morning"]], "response_mode": "offline", "answer": "[{\"word\":\"morning\",\"start\":0.99,\"end\":1.3}]"} {"utterance_id": "1221-135766-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0007.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.82, 4.08, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.82,\"end\":4.08}]"} {"utterance_id": "672-122797-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0041.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen carefully to the entire clip, then report every instance of a time of day as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.71, 1.3, "morning"]], "response_mode": "offline", "answer": "[{\"word\":\"morning\",\"start\":0.71,\"end\":1.3}]"} {"utterance_id": "4992-23283-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0020.flac", "task_type": "lexical_counting", "query": "reason", "prompt": "After the audio ends, tally every instance of the word \"reason\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[4.96, 5.37, "one"], [7.73, 8.12, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.96,\"end\":5.37},{\"word\":\"two\",\"start\":7.73,\"end\":8.12}]"} {"utterance_id": "2300-131720-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0022.flac", "task_type": "lexical_counting", "query": "hopkinson", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"hopkinson\" (word, start, end). If none, reply [].", "labels": [[6.06, 6.77, "one"], [8.08, 8.86, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.06,\"end\":6.77},{\"word\":\"two\",\"start\":8.08,\"end\":8.86}]"} {"utterance_id": "7127-75947-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0013.flac", "task_type": "lexical_detection", "query": "remember", "prompt": "Process the full audio offline. Find every instance of the word \"remember\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.28, 0.7, "remember"]], "response_mode": "offline", "answer": "[{\"word\":\"remember\",\"start\":0.28,\"end\":0.7}]"} {"utterance_id": "3570-5695-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0014.flac", "task_type": "semantic_counting", "query": "money", "prompt": "Listen to the entire clip, then report the count sequence for a money or currency term as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.33, 3.78, "one"], [9.12, 9.79, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.33,\"end\":3.78},{\"word\":\"two\",\"start\":9.12,\"end\":9.79}]"} {"utterance_id": "8455-210777-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0013.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a person's name: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.54, 0.9, "one"], [3.41, 3.77, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.54,\"end\":0.9},{\"word\":\"two\",\"start\":3.41,\"end\":3.77}]"} {"utterance_id": "1221-135767-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0023.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "After listening to the whole recording, output every instance of a state or province name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.35, 10.06, "peninsula"]], "response_mode": "offline", "answer": "[{\"word\":\"peninsula\",\"start\":9.35,\"end\":10.06}]"} {"utterance_id": "6829-68771-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0029.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "After the audio ends, tally every instance of a person's name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.18, 0.5, "one"], [4.47, 4.92, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.18,\"end\":0.5},{\"word\":\"two\",\"start\":4.47,\"end\":4.92}]"} {"utterance_id": "5683-32879-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0019.flac", "task_type": "lexical_detection", "query": "begun", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"begun\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.99, 5.42, "begun"]], "response_mode": "offline", "answer": "[{\"word\":\"begun\",\"start\":4.99,\"end\":5.42}]"} {"utterance_id": "7176-88083-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0002.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Listen to the full audio. Count every occurrence of a body part. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.55, 0.84, "one"], [2.68, 3.07, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.55,\"end\":0.84},{\"word\":\"two\",\"start\":2.68,\"end\":3.07}]"} {"utterance_id": "4446-2271-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0006.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.1, 1.41, "hilda"]], "response_mode": "offline", "answer": "[{\"word\":\"hilda\",\"start\":1.1,\"end\":1.41}]"} {"utterance_id": "4992-41797-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0008.flac", "task_type": "lexical_detection", "query": "contrary", "prompt": "Once the audio has finished, identify every instance of the word \"contrary\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.48, 4.08, "contrary"]], "response_mode": "offline", "answer": "[{\"word\":\"contrary\",\"start\":3.48,\"end\":4.08}]"} {"utterance_id": "61-70968-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0000.flac", "task_type": "lexical_detection", "query": "vanished", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"vanished\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.85, 3.27, "vanished"]], "response_mode": "offline", "answer": "[{\"word\":\"vanished\",\"start\":2.85,\"end\":3.27}]"} {"utterance_id": "4077-13751-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0018.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Once the audio has finished, identify every instance of an organization name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.99, 1.31, "school"], [4.43, 5.37, "albigenses"], [11.51, 12.13, "vaudois"]], "response_mode": "offline", "answer": "[{\"word\":\"school\",\"start\":0.99,\"end\":1.31},{\"word\":\"albigenses\",\"start\":4.43,\"end\":5.37},{\"word\":\"vaudois\",\"start\":11.51,\"end\":12.13}]"} {"utterance_id": "61-70968-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0030.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.78, 4.24, "apprentice"]], "response_mode": "offline", "answer": "[{\"word\":\"apprentice\",\"start\":3.78,\"end\":4.24}]"} {"utterance_id": "3575-170457-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0054.flac", "task_type": "lexical_detection", "query": "fancied", "prompt": "Process the full audio offline. Find every instance of the word \"fancied\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.89, 6.35, "fancied"]], "response_mode": "offline", "answer": "[{\"word\":\"fancied\",\"start\":5.89,\"end\":6.35}]"} {"utterance_id": "1995-1826-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0022.flac", "task_type": "lexical_detection", "query": "early", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"early\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.6, 1.99, "early"]], "response_mode": "offline", "answer": "[{\"word\":\"early\",\"start\":1.6,\"end\":1.99}]"} {"utterance_id": "5683-32866-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0025.flac", "task_type": "lexical_detection", "query": "chimney", "prompt": "Once the audio has finished, identify every instance of the word \"chimney\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.75, 3.2, "chimney"]], "response_mode": "offline", "answer": "[{\"word\":\"chimney\",\"start\":2.75,\"end\":3.2}]"} {"utterance_id": "6829-68769-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0053.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Process the full audio offline. Find every instance of a money or currency term and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.1, 4.49, "payment"], [5.58, 6.15, "dollars"]], "response_mode": "offline", "answer": "[{\"word\":\"payment\",\"start\":4.1,\"end\":4.49},{\"word\":\"dollars\",\"start\":5.58,\"end\":6.15}]"} {"utterance_id": "6829-68771-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.61, 1.83, "beth"]], "response_mode": "offline", "answer": "[{\"word\":\"beth\",\"start\":1.61,\"end\":1.83}]"} {"utterance_id": "1221-135766-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0001.flac", "task_type": "lexical_detection", "query": "bosom", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"bosom\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.62, 9.07, "bosom"]], "response_mode": "offline", "answer": "[{\"word\":\"bosom\",\"start\":8.62,\"end\":9.07}]"} {"utterance_id": "260-123286-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0028.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.74, 1.09, "jaw"], [6.46, 7.0, "teeth"]], "response_mode": "offline", "answer": "[{\"word\":\"jaw\",\"start\":0.74,\"end\":1.09},{\"word\":\"teeth\",\"start\":6.46,\"end\":7.0}]"} {"utterance_id": "5639-40744-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0023.flac", "task_type": "lexical_detection", "query": "bore", "prompt": "After the audio ends, extract all occurrences of the word \"bore\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.75, 8.95, "bore"]], "response_mode": "offline", "answer": "[{\"word\":\"bore\",\"start\":8.75,\"end\":8.95}]"} {"utterance_id": "1188-133604-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0028.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.05, 2.69, "turner's"]], "response_mode": "offline", "answer": "[{\"word\":\"turner's\",\"start\":2.05,\"end\":2.69}]"} {"utterance_id": "237-126133-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0002.flac", "task_type": "lexical_detection", "query": "friends", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"friends\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.9, 8.4, "friends"]], "response_mode": "offline", "answer": "[{\"word\":\"friends\",\"start\":7.9,\"end\":8.4}]"} {"utterance_id": "5683-32865-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.04, 2.64, "wylder"]], "response_mode": "offline", "answer": "[{\"word\":\"wylder\",\"start\":2.04,\"end\":2.64}]"} {"utterance_id": "4992-41797-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0022.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.39, 6.04, "child"]], "response_mode": "offline", "answer": "[{\"word\":\"child\",\"start\":5.39,\"end\":6.04}]"} {"utterance_id": "1580-141084-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0008.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.51, 2.63, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.51,\"end\":2.63}]"} {"utterance_id": "1284-1180-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0011.flac", "task_type": "lexical_detection", "query": "strangers", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"strangers\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.99, 2.55, "strangers"]], "response_mode": "offline", "answer": "[{\"word\":\"strangers\",\"start\":1.99,\"end\":2.55}]"} {"utterance_id": "237-126133-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0006.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.9, 5.62, "phronsie"]], "response_mode": "offline", "answer": "[{\"word\":\"phronsie\",\"start\":4.9,\"end\":5.62}]"} {"utterance_id": "2961-960-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0007.flac", "task_type": "lexical_detection", "query": "plato", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"plato\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.13, 3.74, "plato"]], "response_mode": "offline", "answer": "[{\"word\":\"plato\",\"start\":3.13,\"end\":3.74}]"} {"utterance_id": "1188-133604-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0011.flac", "task_type": "lexical_detection", "query": "dependent", "prompt": "Once the audio has finished, identify every instance of the word \"dependent\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.37, 3.85, "dependent"]], "response_mode": "offline", "answer": "[{\"word\":\"dependent\",\"start\":3.37,\"end\":3.85}]"} {"utterance_id": "237-126133-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0013.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.65, 2.1, "polly"]], "response_mode": "offline", "answer": "[{\"word\":\"polly\",\"start\":1.65,\"end\":2.1}]"} {"utterance_id": "1089-134686-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0010.flac", "task_type": "lexical_detection", "query": "declare", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"declare\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.59, 2.11, "declare"]], "response_mode": "offline", "answer": "[{\"word\":\"declare\",\"start\":1.59,\"end\":2.11}]"} {"utterance_id": "8463-294828-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0003.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Once the audio has finished, identify every instance of a country name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.87, 2.33, "country"]], "response_mode": "offline", "answer": "[{\"word\":\"country\",\"start\":1.87,\"end\":2.33}]"} {"utterance_id": "2961-960-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0009.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.65, 10.02, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":9.65,\"end\":10.02}]"} {"utterance_id": "5639-40744-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0033.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.82, 6.31, "maids"]], "response_mode": "offline", "answer": "[{\"word\":\"maids\",\"start\":5.82,\"end\":6.31}]"} {"utterance_id": "2961-960-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0021.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[13.47, 14.09, "plato's"]], "response_mode": "offline", "answer": "[{\"word\":\"plato's\",\"start\":13.47,\"end\":14.09}]"} {"utterance_id": "121-127105-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0015.flac", "task_type": "lexical_detection", "query": "dropped", "prompt": "Process the full audio offline. Find every instance of the word \"dropped\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.27, 1.56, "dropped"]], "response_mode": "offline", "answer": "[{\"word\":\"dropped\",\"start\":1.27,\"end\":1.56}]"} {"utterance_id": "1089-134691-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0011.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[3.68, 4.25, "faces"], [9.54, 10.04, "hearts"]], "response_mode": "offline", "answer": "[{\"word\":\"faces\",\"start\":3.68,\"end\":4.25},{\"word\":\"hearts\",\"start\":9.54,\"end\":10.04}]"} {"utterance_id": "1284-134647-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0004.flac", "task_type": "lexical_detection", "query": "humanity", "prompt": "After the audio ends, extract all occurrences of the word \"humanity\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.69, 12.38, "humanity"]], "response_mode": "offline", "answer": "[{\"word\":\"humanity\",\"start\":11.69,\"end\":12.38}]"} {"utterance_id": "237-126133-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.21, 1.69, "phronsie"]], "response_mode": "offline", "answer": "[{\"word\":\"phronsie\",\"start\":1.21,\"end\":1.69}]"} {"utterance_id": "8224-274381-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0005.flac", "task_type": "lexical_detection", "query": "hastened", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"hastened\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[12.79, 13.3, "hastened"]], "response_mode": "offline", "answer": "[{\"word\":\"hastened\",\"start\":12.79,\"end\":13.3}]"} {"utterance_id": "260-123286-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0016.flac", "task_type": "lexical_detection", "query": "agitated", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"agitated\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.23, 1.79, "agitated"]], "response_mode": "offline", "answer": "[{\"word\":\"agitated\",\"start\":1.23,\"end\":1.79}]"} {"utterance_id": "260-123288-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0023.flac", "task_type": "lexical_detection", "query": "consent", "prompt": "Once the audio has finished, identify every instance of the word \"consent\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.19, 1.92, "consent"]], "response_mode": "offline", "answer": "[{\"word\":\"consent\",\"start\":1.19,\"end\":1.92}]"} {"utterance_id": "4446-2271-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.22, 1.01, "irene burgoyne"]], "response_mode": "offline", "answer": "[{\"word\":\"irene burgoyne\",\"start\":0.22,\"end\":1.01}]"} {"utterance_id": "5142-33396-0057", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0057.flac", "task_type": "lexical_detection", "query": "ship", "prompt": "After listening to the whole recording, output every instance of the word \"ship\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.71, 1.02, "ship"]], "response_mode": "offline", "answer": "[{\"word\":\"ship\",\"start\":0.71,\"end\":1.02}]"} {"utterance_id": "1089-134691-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0005.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.96, 2.53, "harts"]], "response_mode": "offline", "answer": "[{\"word\":\"harts\",\"start\":1.96,\"end\":2.53}]"} {"utterance_id": "237-126133-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0012.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.89, 4.23, "head"]], "response_mode": "offline", "answer": "[{\"word\":\"head\",\"start\":3.89,\"end\":4.23}]"} {"utterance_id": "1221-135766-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0005.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "After listening to the whole recording, output timed count words for a person's name (one, two, three, ...) as JSON. If none, reply [].", "labels": [[0.51, 0.98, "one"], [10.32, 10.69, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.51,\"end\":0.98},{\"word\":\"two\",\"start\":10.32,\"end\":10.69}]"} {"utterance_id": "6930-75918-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0001.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Do not answer until the audio ends. Then list each instance of a plant name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.71, 3.21, "flowers"]], "response_mode": "offline", "answer": "[{\"word\":\"flowers\",\"start\":2.71,\"end\":3.21}]"} {"utterance_id": "1995-1836-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0006.flac", "task_type": "lexical_detection", "query": "agreeably", "prompt": "After the audio ends, extract all occurrences of the word \"agreeably\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.37, 2.06, "agreeably"]], "response_mode": "offline", "answer": "[{\"word\":\"agreeably\",\"start\":1.37,\"end\":2.06}]"} {"utterance_id": "61-70968-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0026.flac", "task_type": "lexical_detection", "query": "beating", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"beating\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.69, 3.97, "beating"]], "response_mode": "offline", "answer": "[{\"word\":\"beating\",\"start\":3.69,\"end\":3.97}]"} {"utterance_id": "7021-85628-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0003.flac", "task_type": "lexical_detection", "query": "blade", "prompt": "Process the full audio offline. Find every instance of the word \"blade\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.39, 4.72, "blade"]], "response_mode": "offline", "answer": "[{\"word\":\"blade\",\"start\":4.39,\"end\":4.72}]"} {"utterance_id": "672-122797-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0000.flac", "task_type": "lexical_detection", "query": "woods", "prompt": "Process the full audio offline. Find every instance of the word \"woods\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.0, 1.43, "woods"]], "response_mode": "offline", "answer": "[{\"word\":\"woods\",\"start\":1.0,\"end\":1.43}]"} {"utterance_id": "8555-284449-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0017.flac", "task_type": "lexical_detection", "query": "assembled", "prompt": "Process the full audio offline. Find every instance of the word \"assembled\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.64, 6.11, "assembled"]], "response_mode": "offline", "answer": "[{\"word\":\"assembled\",\"start\":5.64,\"end\":6.11}]"} {"utterance_id": "1089-134686-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0006.flac", "task_type": "lexical_detection", "query": "dull", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"dull\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.66, 0.95, "dull"]], "response_mode": "offline", "answer": "[{\"word\":\"dull\",\"start\":0.66,\"end\":0.95}]"} {"utterance_id": "4077-13751-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0004.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After the audio ends, extract all occurrences of an organization name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.75, 10.57, "organizations"], [15.83, 16.27, "church"]], "response_mode": "offline", "answer": "[{\"word\":\"organizations\",\"start\":9.75,\"end\":10.57},{\"word\":\"church\",\"start\":15.83,\"end\":16.27}]"} {"utterance_id": "8224-274381-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0002.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the full audio. After it finishes, find every instance of an organization name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.7, 4.25, "scottish covenanters"], [7.25, 8.3, "english parliament"]], "response_mode": "offline", "answer": "[{\"word\":\"scottish covenanters\",\"start\":2.7,\"end\":4.25},{\"word\":\"english parliament\",\"start\":7.25,\"end\":8.3}]"} {"utterance_id": "3575-170457-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0026.flac", "task_type": "lexical_detection", "query": "wasted", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"wasted\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.06, 11.61, "wasted"]], "response_mode": "offline", "answer": "[{\"word\":\"wasted\",\"start\":11.06,\"end\":11.61}]"} {"utterance_id": "1320-122617-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0002.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Do not answer until the audio ends. Then list each instance of an animal name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.99, 4.32, "ass"], [7.34, 7.71, "bear's"]], "response_mode": "offline", "answer": "[{\"word\":\"ass\",\"start\":3.99,\"end\":4.32},{\"word\":\"bear's\",\"start\":7.34,\"end\":7.71}]"} {"utterance_id": "5142-33396-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0012.flac", "task_type": "lexical_detection", "query": "winter", "prompt": "After the audio ends, extract all occurrences of the word \"winter\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.74, 2.02, "winter"]], "response_mode": "offline", "answer": "[{\"word\":\"winter\",\"start\":1.74,\"end\":2.02}]"} {"utterance_id": "6930-76324-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0024.flac", "task_type": "lexical_detection", "query": "candle", "prompt": "When the utterance is complete, report all instances of the word \"candle\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.72, 2.14, "candle"]], "response_mode": "offline", "answer": "[{\"word\":\"candle\",\"start\":1.72,\"end\":2.14}]"} {"utterance_id": "1580-141083-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0038.flac", "task_type": "lexical_detection", "query": "habit", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"habit\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.97, 4.38, "habit"]], "response_mode": "offline", "answer": "[{\"word\":\"habit\",\"start\":3.97,\"end\":4.38}]"} {"utterance_id": "1221-135766-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0003.flac", "task_type": "lexical_detection", "query": "precisely", "prompt": "Once the audio has finished, identify every instance of the word \"precisely\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[11.47, 12.28, "precisely"]], "response_mode": "offline", "answer": "[{\"word\":\"precisely\",\"start\":11.47,\"end\":12.28}]"} {"utterance_id": "237-134500-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0002.flac", "task_type": "lexical_detection", "query": "clouds", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"clouds\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.26, 3.71, "clouds"]], "response_mode": "offline", "answer": "[{\"word\":\"clouds\",\"start\":3.26,\"end\":3.71}]"} {"utterance_id": "4970-29095-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.54, 1.54, "margaret bolton"]], "response_mode": "offline", "answer": "[{\"word\":\"margaret bolton\",\"start\":0.54,\"end\":1.54}]"} {"utterance_id": "5142-36377-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0014.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Listen to the entire clip, then report the count sequence for a body part as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.71, 7.07, "one"], [8.57, 9.13, "two"], [10.19, 10.49, "three"], [24.9, 25.31, "four"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.71,\"end\":7.07},{\"word\":\"two\",\"start\":8.57,\"end\":9.13},{\"word\":\"three\",\"start\":10.19,\"end\":10.49},{\"word\":\"four\",\"start\":24.9,\"end\":25.31}]"} {"utterance_id": "5683-32879-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0016.flac", "task_type": "semantic_detection", "query": "color", "prompt": "You will hear a complete utterance. When it ends, list every instance of a color name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.78, 1.2, "wan"]], "response_mode": "offline", "answer": "[{\"word\":\"wan\",\"start\":0.78,\"end\":1.2}]"} {"utterance_id": "908-31957-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0012.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Once the audio has finished, list each occurrence of a number as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[1.5, 1.85, "one"], [3.09, 3.39, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.5,\"end\":1.85},{\"word\":\"two\",\"start\":3.09,\"end\":3.39}]"} {"utterance_id": "672-122797-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0034.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "You will hear a complete utterance. When it ends, list every instance of a plant name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.21, 4.63, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":4.21,\"end\":4.63}]"} {"utterance_id": "3570-5695-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0001.flac", "task_type": "lexical_detection", "query": "expected", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"expected\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.94, 6.52, "expected"]], "response_mode": "offline", "answer": "[{\"word\":\"expected\",\"start\":5.94,\"end\":6.52}]"} {"utterance_id": "8455-210777-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0025.flac", "task_type": "lexical_detection", "query": "die", "prompt": "Once the audio has finished, identify every instance of the word \"die\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.92, 3.33, "die"]], "response_mode": "offline", "answer": "[{\"word\":\"die\",\"start\":2.92,\"end\":3.33}]"} {"utterance_id": "672-122797-0070", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0070.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.56, 3.97, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":3.56,\"end\":3.97}]"} {"utterance_id": "5639-40744-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0006.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.32, 6.65, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":6.32,\"end\":6.65}]"} {"utterance_id": "3570-5696-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0005.flac", "task_type": "lexical_counting", "query": "expenditure", "prompt": "When the utterance is complete, report how the word \"expenditure\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.38, 3.13, "one"], [11.13, 11.79, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.38,\"end\":3.13},{\"word\":\"two\",\"start\":11.13,\"end\":11.79}]"} {"utterance_id": "5683-32879-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0011.flac", "task_type": "lexical_detection", "query": "partook", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"partook\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.68, 3.18, "partook"]], "response_mode": "offline", "answer": "[{\"word\":\"partook\",\"start\":2.68,\"end\":3.18}]"} {"utterance_id": "2830-3980-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0048.flac", "task_type": "lexical_detection", "query": "suffered", "prompt": "After listening to the whole recording, output every instance of the word \"suffered\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.81, 3.48, "suffered"]], "response_mode": "offline", "answer": "[{\"word\":\"suffered\",\"start\":2.81,\"end\":3.48}]"} {"utterance_id": "2961-961-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0012.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Do not answer until the audio ends. Then list each instance of a city name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.51, 3.99, "athens"], [5.61, 6.21, "cities"]], "response_mode": "offline", "answer": "[{\"word\":\"athens\",\"start\":3.51,\"end\":3.99},{\"word\":\"cities\",\"start\":5.61,\"end\":6.21}]"} {"utterance_id": "1320-122612-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0011.flac", "task_type": "lexical_detection", "query": "links", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"links\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.65, 5.05, "links"]], "response_mode": "offline", "answer": "[{\"word\":\"links\",\"start\":4.65,\"end\":5.05}]"} {"utterance_id": "4077-13751-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0014.flac", "task_type": "lexical_counting", "query": "county", "prompt": "After the audio ends, tally every instance of the word \"county\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[5.9, 6.36, "one"], [9.32, 9.75, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.9,\"end\":6.36},{\"word\":\"two\",\"start\":9.32,\"end\":9.75}]"} {"utterance_id": "2961-961-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0017.flac", "task_type": "lexical_detection", "query": "pursuit", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"pursuit\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.13, 3.72, "pursuit"]], "response_mode": "offline", "answer": "[{\"word\":\"pursuit\",\"start\":3.13,\"end\":3.72}]"} {"utterance_id": "7021-79740-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0007.flac", "task_type": "lexical_detection", "query": "altered", "prompt": "After the audio ends, extract all occurrences of the word \"altered\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.92, 4.31, "altered"]], "response_mode": "offline", "answer": "[{\"word\":\"altered\",\"start\":3.92,\"end\":4.31}]"} {"utterance_id": "4992-41797-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0019.flac", "task_type": "semantic_detection", "query": "color", "prompt": "When the utterance is complete, report all instances of a color name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.66, 2.01, "black"], [3.72, 4.11, "brown"]], "response_mode": "offline", "answer": "[{\"word\":\"black\",\"start\":1.66,\"end\":2.01},{\"word\":\"brown\",\"start\":3.72,\"end\":4.11}]"} {"utterance_id": "1580-141084-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0050.flac", "task_type": "lexical_detection", "query": "game", "prompt": "Process the full audio offline. Find every instance of the word \"game\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.84, 2.06, "game"]], "response_mode": "offline", "answer": "[{\"word\":\"game\",\"start\":1.84,\"end\":2.06}]"} {"utterance_id": "6829-68769-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0048.flac", "task_type": "lexical_detection", "query": "proceedings", "prompt": "Once the audio has finished, identify every instance of the word \"proceedings\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.42, 6.18, "proceedings"]], "response_mode": "offline", "answer": "[{\"word\":\"proceedings\",\"start\":5.42,\"end\":6.18}]"} {"utterance_id": "1284-1181-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0017.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Do not answer until the audio ends. Then list each instance of a country name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.88, 1.46, "oz"]], "response_mode": "offline", "answer": "[{\"word\":\"oz\",\"start\":0.88,\"end\":1.46}]"} {"utterance_id": "5142-36377-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0020.flac", "task_type": "lexical_detection", "query": "party", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"party\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.58, 3.04, "party"]], "response_mode": "offline", "answer": "[{\"word\":\"party\",\"start\":2.58,\"end\":3.04}]"} {"utterance_id": "5683-32879-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0025.flac", "task_type": "lexical_detection", "query": "dorcas", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"dorcas\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.91, 1.33, "dorcas"]], "response_mode": "offline", "answer": "[{\"word\":\"dorcas\",\"start\":0.91,\"end\":1.33}]"} {"utterance_id": "6829-68771-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0024.flac", "task_type": "lexical_detection", "query": "wandered", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"wandered\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.07, 4.48, "wandered"]], "response_mode": "offline", "answer": "[{\"word\":\"wandered\",\"start\":4.07,\"end\":4.48}]"} {"utterance_id": "6930-75918-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0001.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Do not answer until the audio ends. Then list each instance of a food name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[11.03, 11.56, "supper"]], "response_mode": "offline", "answer": "[{\"word\":\"supper\",\"start\":11.03,\"end\":11.56}]"} {"utterance_id": "7176-92135-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0042.flac", "task_type": "lexical_detection", "query": "stage", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"stage\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.37, 4.74, "stage"]], "response_mode": "offline", "answer": "[{\"word\":\"stage\",\"start\":4.37,\"end\":4.74}]"} {"utterance_id": "2300-131720-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0032.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.32, 1.84, "edison"]], "response_mode": "offline", "answer": "[{\"word\":\"edison\",\"start\":1.32,\"end\":1.84}]"} {"utterance_id": "61-70968-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0045.flac", "task_type": "lexical_detection", "query": "men", "prompt": "Once the audio has finished, identify every instance of the word \"men\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.66, 3.05, "men"]], "response_mode": "offline", "answer": "[{\"word\":\"men\",\"start\":2.66,\"end\":3.05}]"} {"utterance_id": "8463-294828-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0027.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.18, 4.63, "plants"]], "response_mode": "offline", "answer": "[{\"word\":\"plants\",\"start\":4.18,\"end\":4.63}]"} {"utterance_id": "2300-131720-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0027.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen carefully to the entire clip, then report every instance of a food name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.13, 5.57, "water"]], "response_mode": "offline", "answer": "[{\"word\":\"water\",\"start\":5.13,\"end\":5.57}]"} {"utterance_id": "3575-170457-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0041.flac", "task_type": "lexical_detection", "query": "errand", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"errand\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.42, 2.79, "errand"]], "response_mode": "offline", "answer": "[{\"word\":\"errand\",\"start\":2.42,\"end\":2.79}]"} {"utterance_id": "2094-142345-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0000.flac", "task_type": "semantic_counting", "query": "color", "prompt": "After the audio ends, tally every instance of a color name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[2.38, 2.66, "one"], [4.6, 5.24, "two"], [12.48, 12.81, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.38,\"end\":2.66},{\"word\":\"two\",\"start\":4.6,\"end\":5.24},{\"word\":\"three\",\"start\":12.48,\"end\":12.81}]"} {"utterance_id": "4507-16021-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0024.flac", "task_type": "lexical_detection", "query": "each", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"each\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.97, 4.27, "each"]], "response_mode": "offline", "answer": "[{\"word\":\"each\",\"start\":3.97,\"end\":4.27}]"} {"utterance_id": "2094-142345-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0035.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.93, 2.33, "aunt"]], "response_mode": "offline", "answer": "[{\"word\":\"aunt\",\"start\":1.93,\"end\":2.33}]"} {"utterance_id": "2830-3980-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0031.flac", "task_type": "lexical_detection", "query": "provinces", "prompt": "Process the full audio offline. Find every instance of the word \"provinces\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.96, 2.54, "provinces"]], "response_mode": "offline", "answer": "[{\"word\":\"provinces\",\"start\":1.96,\"end\":2.54}]"} {"utterance_id": "237-134500-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0025.flac", "task_type": "lexical_detection", "query": "emil", "prompt": "After listening to the whole recording, output every instance of the word \"emil\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.8, 1.39, "emil"]], "response_mode": "offline", "answer": "[{\"word\":\"emil\",\"start\":0.8,\"end\":1.39}]"} {"utterance_id": "6930-75918-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0005.flac", "task_type": "lexical_detection", "query": "chest", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"chest\" with fields word, start, end (seconds). If none, reply [].", "labels": [[10.71, 11.26, "chest"]], "response_mode": "offline", "answer": "[{\"word\":\"chest\",\"start\":10.71,\"end\":11.26}]"} {"utterance_id": "4446-2275-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0003.flac", "task_type": "lexical_detection", "query": "empty", "prompt": "When the utterance is complete, report all instances of the word \"empty\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.88, 1.21, "empty"]], "response_mode": "offline", "answer": "[{\"word\":\"empty\",\"start\":0.88,\"end\":1.21}]"} {"utterance_id": "1284-1180-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0008.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "You will hear a complete utterance. When it ends, list every instance of a plant name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.51, 5.77, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":5.51,\"end\":5.77}]"} {"utterance_id": "1188-133604-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0012.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.51, 2.36, "colorist"], [5.2, 5.57, "singer"]], "response_mode": "offline", "answer": "[{\"word\":\"colorist\",\"start\":1.51,\"end\":2.36},{\"word\":\"singer\",\"start\":5.2,\"end\":5.57}]"} {"utterance_id": "4446-2273-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0004.flac", "task_type": "lexical_detection", "query": "cookery", "prompt": "After the audio ends, extract all occurrences of the word \"cookery\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.45, 4.97, "cookery"]], "response_mode": "offline", "answer": "[{\"word\":\"cookery\",\"start\":4.45,\"end\":4.97}]"} {"utterance_id": "8555-284449-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0020.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Listen to the full audio. After it finishes, find every instance of a country name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.76, 2.27, "countries"]], "response_mode": "offline", "answer": "[{\"word\":\"countries\",\"start\":1.76,\"end\":2.27}]"} {"utterance_id": "5142-33396-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0016.flac", "task_type": "lexical_detection", "query": "coast", "prompt": "Process the full audio offline. Find every instance of the word \"coast\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.01, 1.35, "coast"]], "response_mode": "offline", "answer": "[{\"word\":\"coast\",\"start\":1.01,\"end\":1.35}]"} {"utterance_id": "8455-210777-0068", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0068.flac", "task_type": "lexical_detection", "query": "sufficient", "prompt": "Process the full audio offline. Find every instance of the word \"sufficient\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.22, 1.7, "sufficient"]], "response_mode": "offline", "answer": "[{\"word\":\"sufficient\",\"start\":1.22,\"end\":1.7}]"} {"utterance_id": "4446-2275-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0010.flac", "task_type": "lexical_detection", "query": "blaze", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"blaze\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.98, 3.54, "blaze"]], "response_mode": "offline", "answer": "[{\"word\":\"blaze\",\"start\":2.98,\"end\":3.54}]"} {"utterance_id": "6829-68771-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0011.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Process the full audio offline. Find every instance of a food name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.6, 4.23, "repast"]], "response_mode": "offline", "answer": "[{\"word\":\"repast\",\"start\":3.6,\"end\":4.23}]"} {"utterance_id": "8224-274381-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0012.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "When the utterance is complete, report all instances of an organization name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[18.66, 19.47, "royalists"]], "response_mode": "offline", "answer": "[{\"word\":\"royalists\",\"start\":18.66,\"end\":19.47}]"} {"utterance_id": "8224-274384-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0004.flac", "task_type": "lexical_counting", "query": "king", "prompt": "Listen to the entire clip, then report the count sequence for the word \"king\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.4, 6.87, "one"], [19.57, 20.14, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.4,\"end\":6.87},{\"word\":\"two\",\"start\":19.57,\"end\":20.14}]"} {"utterance_id": "5142-33396-0063", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0063.flac", "task_type": "lexical_detection", "query": "gift", "prompt": "After listening to the whole recording, output every instance of the word \"gift\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.22, 1.47, "gift"]], "response_mode": "offline", "answer": "[{\"word\":\"gift\",\"start\":1.22,\"end\":1.47}]"} {"utterance_id": "2961-961-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0012.flac", "task_type": "lexical_counting", "query": "best", "prompt": "Process the full audio offline. For every occurrence of the word \"best\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.1, 5.51, "one"], [9.0, 9.51, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.1,\"end\":5.51},{\"word\":\"two\",\"start\":9.0,\"end\":9.51}]"} {"utterance_id": "672-122797-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0008.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an animal name with fields word, start, end (seconds). If none, reply [].", "labels": [[27.79, 28.38, "horses"]], "response_mode": "offline", "answer": "[{\"word\":\"horses\",\"start\":27.79,\"end\":28.38}]"} {"utterance_id": "3570-5696-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0001.flac", "task_type": "lexical_detection", "query": "start", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"start\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[6.26, 6.7, "start"]], "response_mode": "offline", "answer": "[{\"word\":\"start\",\"start\":6.26,\"end\":6.7}]"} {"utterance_id": "3575-170457-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0017.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.71, 3.26, "adviser"]], "response_mode": "offline", "answer": "[{\"word\":\"adviser\",\"start\":2.71,\"end\":3.26}]"} {"utterance_id": "4507-16021-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0002.flac", "task_type": "lexical_detection", "query": "thus", "prompt": "After listening to the whole recording, output every instance of the word \"thus\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.28, 0.89, "thus"]], "response_mode": "offline", "answer": "[{\"word\":\"thus\",\"start\":0.28,\"end\":0.89}]"} {"utterance_id": "7127-75947-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0040.flac", "task_type": "lexical_detection", "query": "inhabitants", "prompt": "After listening to the whole recording, output every instance of the word \"inhabitants\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.0, 5.69, "inhabitants"]], "response_mode": "offline", "answer": "[{\"word\":\"inhabitants\",\"start\":5.0,\"end\":5.69}]"} {"utterance_id": "8230-279154-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0018.flac", "task_type": "lexical_detection", "query": "temporal", "prompt": "When the utterance is complete, report all instances of the word \"temporal\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.83, 4.44, "temporal"]], "response_mode": "offline", "answer": "[{\"word\":\"temporal\",\"start\":3.83,\"end\":4.44}]"} {"utterance_id": "61-70968-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0002.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After the audio ends, extract all occurrences of a color name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.29, 0.8, "golden"]], "response_mode": "offline", "answer": "[{\"word\":\"golden\",\"start\":0.29,\"end\":0.8}]"} {"utterance_id": "7021-85628-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.99, 2.55, "anders"]], "response_mode": "offline", "answer": "[{\"word\":\"anders\",\"start\":1.99,\"end\":2.55}]"} {"utterance_id": "8463-294825-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0009.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.4, 4.94, "fighter"]], "response_mode": "offline", "answer": "[{\"word\":\"fighter\",\"start\":4.4,\"end\":4.94}]"} {"utterance_id": "260-123288-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0024.flac", "task_type": "lexical_detection", "query": "inch", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"inch\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.94, 5.22, "inch"]], "response_mode": "offline", "answer": "[{\"word\":\"inch\",\"start\":4.94,\"end\":5.22}]"} {"utterance_id": "2300-131720-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0015.flac", "task_type": "lexical_detection", "query": "desired", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"desired\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.41, 1.92, "desired"]], "response_mode": "offline", "answer": "[{\"word\":\"desired\",\"start\":1.41,\"end\":1.92}]"} {"utterance_id": "8224-274384-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0009.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, list every instance of an organization name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.6, 1.3, "parliament"]], "response_mode": "offline", "answer": "[{\"word\":\"parliament\",\"start\":0.6,\"end\":1.3}]"} {"utterance_id": "3575-170457-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0018.flac", "task_type": "lexical_detection", "query": "sixteenth", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"sixteenth\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.81, 2.49, "sixteenth"]], "response_mode": "offline", "answer": "[{\"word\":\"sixteenth\",\"start\":1.81,\"end\":2.49}]"} {"utterance_id": "1995-1837-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0010.flac", "task_type": "lexical_detection", "query": "weeping", "prompt": "Once the audio has finished, identify every instance of the word \"weeping\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.75, 3.29, "weeping"]], "response_mode": "offline", "answer": "[{\"word\":\"weeping\",\"start\":2.75,\"end\":3.29}]"} {"utterance_id": "1089-134686-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.97, 1.43, "ennis"]], "response_mode": "offline", "answer": "[{\"word\":\"ennis\",\"start\":0.97,\"end\":1.43}]"} {"utterance_id": "61-70968-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0053.flac", "task_type": "lexical_detection", "query": "esquire", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"esquire\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.91, 1.56, "esquire"]], "response_mode": "offline", "answer": "[{\"word\":\"esquire\",\"start\":0.91,\"end\":1.56}]"} {"utterance_id": "237-134500-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0013.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.35, 0.78, "marie"]], "response_mode": "offline", "answer": "[{\"word\":\"marie\",\"start\":0.35,\"end\":0.78}]"} {"utterance_id": "2961-961-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0018.flac", "task_type": "lexical_detection", "query": "famous", "prompt": "After the audio ends, extract all occurrences of the word \"famous\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.96, 1.41, "famous"]], "response_mode": "offline", "answer": "[{\"word\":\"famous\",\"start\":0.96,\"end\":1.41}]"} {"utterance_id": "6829-68771-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0003.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the full audio. After it finishes, find every instance of an organization name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.63, 1.77, "democratic committee"]], "response_mode": "offline", "answer": "[{\"word\":\"democratic committee\",\"start\":0.63,\"end\":1.77}]"} {"utterance_id": "672-122797-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0008.flac", "task_type": "lexical_counting", "query": "trees", "prompt": "After the audio ends, tally every instance of the word \"trees\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[11.02, 11.66, "one"], [17.78, 18.37, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":11.02,\"end\":11.66},{\"word\":\"two\",\"start\":17.78,\"end\":18.37}]"} {"utterance_id": "4446-2275-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0020.flac", "task_type": "lexical_detection", "query": "myself", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"myself\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.85, 1.35, "myself"]], "response_mode": "offline", "answer": "[{\"word\":\"myself\",\"start\":0.85,\"end\":1.35}]"} {"utterance_id": "1188-133604-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0032.flac", "task_type": "lexical_counting", "query": "psyche", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"psyche\" with word/start/end. If none, reply [].", "labels": [[5.38, 6.01, "one"], [8.11, 8.68, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.38,\"end\":6.01},{\"word\":\"two\",\"start\":8.11,\"end\":8.68}]"} {"utterance_id": "1221-135766-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0006.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "After listening to the whole recording, output every instance of a time of day with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.71, 2.1, "morning"]], "response_mode": "offline", "answer": "[{\"word\":\"morning\",\"start\":1.71,\"end\":2.1}]"} {"utterance_id": "7127-75946-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0024.flac", "task_type": "lexical_detection", "query": "understand", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"understand\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.69, 3.35, "understand"]], "response_mode": "offline", "answer": "[{\"word\":\"understand\",\"start\":2.69,\"end\":3.35}]"} {"utterance_id": "1995-1837-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0016.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.33, 0.55, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.33,\"end\":0.55}]"} {"utterance_id": "2830-3980-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0017.flac", "task_type": "lexical_detection", "query": "call", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"call\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.04, 3.41, "call"]], "response_mode": "offline", "answer": "[{\"word\":\"call\",\"start\":3.04,\"end\":3.41}]"} {"utterance_id": "260-123286-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0031.flac", "task_type": "lexical_detection", "query": "cavern", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"cavern\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.28, 5.04, "cavern"]], "response_mode": "offline", "answer": "[{\"word\":\"cavern\",\"start\":4.28,\"end\":5.04}]"} {"utterance_id": "1580-141084-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0010.flac", "task_type": "lexical_detection", "query": "bye", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"bye\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.71, 3.97, "bye"]], "response_mode": "offline", "answer": "[{\"word\":\"bye\",\"start\":3.71,\"end\":3.97}]"} {"utterance_id": "908-31957-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0022.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Process the full audio offline. Find every instance of a plant name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.5, 11.14, "leaves"]], "response_mode": "offline", "answer": "[{\"word\":\"leaves\",\"start\":10.5,\"end\":11.14}]"} {"utterance_id": "8555-292519-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0007.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "When the utterance is complete, report all instances of a plant name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.82, 6.39, "flowers"]], "response_mode": "offline", "answer": "[{\"word\":\"flowers\",\"start\":5.82,\"end\":6.39}]"} {"utterance_id": "8224-274381-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0004.flac", "task_type": "semantic_counting", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, produce a running count of an organization name: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[3.58, 4.53, "one"], [9.58, 10.06, "two"], [14.72, 15.14, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.58,\"end\":4.53},{\"word\":\"two\",\"start\":9.58,\"end\":10.06},{\"word\":\"three\",\"start\":14.72,\"end\":15.14}]"} {"utterance_id": "2961-961-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0014.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.46, 1.55, "nine thousand"], [5.11, 6.03, "eight thousand"]], "response_mode": "offline", "answer": "[{\"word\":\"nine thousand\",\"start\":0.46,\"end\":1.55},{\"word\":\"eight thousand\",\"start\":5.11,\"end\":6.03}]"} {"utterance_id": "4970-29095-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0036.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.04, 2.53, "career"], [3.67, 4.06, "career"]], "response_mode": "offline", "answer": "[{\"word\":\"career\",\"start\":2.04,\"end\":2.53},{\"word\":\"career\",\"start\":3.67,\"end\":4.06}]"} {"utterance_id": "121-127105-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0025.flac", "task_type": "lexical_detection", "query": "departing", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"departing\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.43, 1.0, "departing"]], "response_mode": "offline", "answer": "[{\"word\":\"departing\",\"start\":0.43,\"end\":1.0}]"} {"utterance_id": "6930-76324-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0019.flac", "task_type": "lexical_detection", "query": "dust", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"dust\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.61, 0.91, "dust"]], "response_mode": "offline", "answer": "[{\"word\":\"dust\",\"start\":0.61,\"end\":0.91}]"} {"utterance_id": "5105-28241-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0015.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.44, 3.31, "lieutenant procope"]], "response_mode": "offline", "answer": "[{\"word\":\"lieutenant procope\",\"start\":2.44,\"end\":3.31}]"} {"utterance_id": "1995-1836-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0011.flac", "task_type": "lexical_detection", "query": "positively", "prompt": "Once the audio has finished, identify every instance of the word \"positively\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.11, 0.9, "positively"]], "response_mode": "offline", "answer": "[{\"word\":\"positively\",\"start\":0.11,\"end\":0.9}]"} {"utterance_id": "8463-287645-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0006.flac", "task_type": "lexical_detection", "query": "slipped", "prompt": "Once the audio has finished, identify every instance of the word \"slipped\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.31, 4.77, "slipped"]], "response_mode": "offline", "answer": "[{\"word\":\"slipped\",\"start\":4.31,\"end\":4.77}]"} {"utterance_id": "4446-2271-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0001.flac", "task_type": "lexical_detection", "query": "americans", "prompt": "Process the full audio offline. Find every instance of the word \"americans\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.45, 4.07, "americans"]], "response_mode": "offline", "answer": "[{\"word\":\"americans\",\"start\":3.45,\"end\":4.07}]"} {"utterance_id": "908-157963-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0022.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.69, 4.2, "queen"]], "response_mode": "offline", "answer": "[{\"word\":\"queen\",\"start\":3.69,\"end\":4.2}]"} {"utterance_id": "2094-142345-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0051.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "After listening to the whole recording, output every instance of an animal name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.5, 2.97, "whelps"]], "response_mode": "offline", "answer": "[{\"word\":\"whelps\",\"start\":2.5,\"end\":2.97}]"} {"utterance_id": "4992-23283-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0013.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.68, 2.34, "miss milner's"]], "response_mode": "offline", "answer": "[{\"word\":\"miss milner's\",\"start\":1.68,\"end\":2.34}]"} {"utterance_id": "1284-1181-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0012.flac", "task_type": "lexical_detection", "query": "powder", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"powder\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.91, 3.29, "powder"]], "response_mode": "offline", "answer": "[{\"word\":\"powder\",\"start\":2.91,\"end\":3.29}]"} {"utterance_id": "2830-3979-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0001.flac", "task_type": "lexical_counting", "query": "luther", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"luther\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[2.62, 2.93, "one"], [11.76, 12.16, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.62,\"end\":2.93},{\"word\":\"two\",\"start\":11.76,\"end\":12.16}]"} {"utterance_id": "7176-92135-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0022.flac", "task_type": "lexical_detection", "query": "nobler", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"nobler\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.84, 3.36, "nobler"]], "response_mode": "offline", "answer": "[{\"word\":\"nobler\",\"start\":2.84,\"end\":3.36}]"} {"utterance_id": "5683-32866-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0009.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "After the audio ends, extract all occurrences of a state or province name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.26, 12.84, "county"]], "response_mode": "offline", "answer": "[{\"word\":\"county\",\"start\":12.26,\"end\":12.84}]"} {"utterance_id": "1580-141083-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0000.flac", "task_type": "lexical_detection", "query": "avoid", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"avoid\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.05, 2.41, "avoid"]], "response_mode": "offline", "answer": "[{\"word\":\"avoid\",\"start\":2.05,\"end\":2.41}]"} {"utterance_id": "7127-75946-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0029.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a profession: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.62, 1.08, "one"], [4.8, 5.28, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.62,\"end\":1.08},{\"word\":\"two\",\"start\":4.8,\"end\":5.28}]"} {"utterance_id": "1580-141084-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0018.flac", "task_type": "lexical_detection", "query": "conclusion", "prompt": "When the utterance is complete, report all instances of the word \"conclusion\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.97, "conclusion"]], "response_mode": "offline", "answer": "[{\"word\":\"conclusion\",\"start\":2.35,\"end\":2.97}]"} {"utterance_id": "1580-141083-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0047.flac", "task_type": "lexical_detection", "query": "papers", "prompt": "After listening to the whole recording, output every instance of the word \"papers\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.25, 1.69, "papers"]], "response_mode": "offline", "answer": "[{\"word\":\"papers\",\"start\":1.25,\"end\":1.69}]"} {"utterance_id": "61-70968-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0048.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.35, 0.68, "henry"]], "response_mode": "offline", "answer": "[{\"word\":\"henry\",\"start\":0.35,\"end\":0.68}]"} {"utterance_id": "4992-23283-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.8, 2.36, "miss woodley"]], "response_mode": "offline", "answer": "[{\"word\":\"miss woodley\",\"start\":1.8,\"end\":2.36}]"} {"utterance_id": "6829-68771-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.99, 6.74, "mister forbes"]], "response_mode": "offline", "answer": "[{\"word\":\"mister forbes\",\"start\":5.99,\"end\":6.74}]"} {"utterance_id": "8455-210777-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0013.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.54, 0.9, "jack"], [3.41, 3.77, "eva"]], "response_mode": "offline", "answer": "[{\"word\":\"jack\",\"start\":0.54,\"end\":0.9},{\"word\":\"eva\",\"start\":3.41,\"end\":3.77}]"} {"utterance_id": "6930-81414-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0022.flac", "task_type": "lexical_detection", "query": "acting", "prompt": "When the utterance is complete, report all instances of the word \"acting\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.15, 1.62, "acting"]], "response_mode": "offline", "answer": "[{\"word\":\"acting\",\"start\":1.15,\"end\":1.62}]"} {"utterance_id": "2094-142345-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0004.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.47, 1.69, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":1.47,\"end\":1.69}]"} {"utterance_id": "1580-141083-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0005.flac", "task_type": "lexical_detection", "query": "absent", "prompt": "When the utterance is complete, report all instances of the word \"absent\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 1.04, "absent"]], "response_mode": "offline", "answer": "[{\"word\":\"absent\",\"start\":0.53,\"end\":1.04}]"} {"utterance_id": "1580-141083-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0037.flac", "task_type": "lexical_detection", "query": "bedroom", "prompt": "After listening to the whole recording, output every instance of the word \"bedroom\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.96, 4.36, "bedroom"]], "response_mode": "offline", "answer": "[{\"word\":\"bedroom\",\"start\":3.96,\"end\":4.36}]"} {"utterance_id": "6930-76324-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0016.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Once the audio has finished, identify every instance of an animal name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.41, 5.91, "kitten"]], "response_mode": "offline", "answer": "[{\"word\":\"kitten\",\"start\":5.41,\"end\":5.91}]"} {"utterance_id": "6829-68771-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0034.flac", "task_type": "lexical_detection", "query": "wish", "prompt": "After the audio ends, extract all occurrences of the word \"wish\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.41, 0.69, "wish"]], "response_mode": "offline", "answer": "[{\"word\":\"wish\",\"start\":0.41,\"end\":0.69}]"} {"utterance_id": "8463-294828-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0035.flac", "task_type": "lexical_detection", "query": "missed", "prompt": "After listening to the whole recording, output every instance of the word \"missed\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.52, 6.85, "missed"]], "response_mode": "offline", "answer": "[{\"word\":\"missed\",\"start\":6.52,\"end\":6.85}]"} {"utterance_id": "2961-960-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0008.flac", "task_type": "lexical_detection", "query": "arranged", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"arranged\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.16, 3.75, "arranged"]], "response_mode": "offline", "answer": "[{\"word\":\"arranged\",\"start\":3.16,\"end\":3.75}]"} {"utterance_id": "7021-79740-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0008.flac", "task_type": "lexical_counting", "query": "piazza", "prompt": "Once the audio has finished, list each occurrence of the word \"piazza\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[4.89, 5.39, "one"], [13.03, 13.61, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.89,\"end\":5.39},{\"word\":\"two\",\"start\":13.03,\"end\":13.61}]"} {"utterance_id": "61-70968-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0029.flac", "task_type": "lexical_detection", "query": "helped", "prompt": "Once the audio has finished, identify every instance of the word \"helped\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.81, 1.04, "helped"]], "response_mode": "offline", "answer": "[{\"word\":\"helped\",\"start\":0.81,\"end\":1.04}]"} {"utterance_id": "4970-29095-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0024.flac", "task_type": "lexical_detection", "query": "occupied", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"occupied\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.69, 1.26, "occupied"]], "response_mode": "offline", "answer": "[{\"word\":\"occupied\",\"start\":0.69,\"end\":1.26}]"} {"utterance_id": "1221-135767-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0021.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a plant name (word, start, end). If none, reply [].", "labels": [[8.17, 8.63, "one"], [11.93, 12.44, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":8.17,\"end\":8.63},{\"word\":\"two\",\"start\":11.93,\"end\":12.44}]"} {"utterance_id": "7176-88083-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0022.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an ordinal number with fields word, start, end (seconds). If none, reply [].", "labels": [[2.32, 2.67, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":2.32,\"end\":2.67}]"} {"utterance_id": "7127-75947-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0039.flac", "task_type": "lexical_detection", "query": "hard", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"hard\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.63, 6.04, "hard"]], "response_mode": "offline", "answer": "[{\"word\":\"hard\",\"start\":5.63,\"end\":6.04}]"} {"utterance_id": "5142-33396-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0025.flac", "task_type": "lexical_detection", "query": "obeyed", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"obeyed\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.84, 2.37, "obeyed"]], "response_mode": "offline", "answer": "[{\"word\":\"obeyed\",\"start\":1.84,\"end\":2.37}]"} {"utterance_id": "5683-32879-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0023.flac", "task_type": "lexical_detection", "query": "generous", "prompt": "After listening to the whole recording, output every instance of the word \"generous\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.11, 4.74, "generous"]], "response_mode": "offline", "answer": "[{\"word\":\"generous\",\"start\":4.11,\"end\":4.74}]"} {"utterance_id": "7127-75946-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0000.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "You will hear a complete utterance. When it ends, list every instance of a time of day you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.1, 3.86, "five o'clock"]], "response_mode": "offline", "answer": "[{\"word\":\"five o'clock\",\"start\":3.1,\"end\":3.86}]"} {"utterance_id": "2830-3980-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0029.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a family relation with fields word, start, end (seconds). If none, reply [].", "labels": [[1.01, 1.39, "brethren"]], "response_mode": "offline", "answer": "[{\"word\":\"brethren\",\"start\":1.01,\"end\":1.39}]"} {"utterance_id": "7127-75946-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0014.flac", "task_type": "lexical_detection", "query": "elegance", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"elegance\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.45, 8.89, "elegance"]], "response_mode": "offline", "answer": "[{\"word\":\"elegance\",\"start\":8.45,\"end\":8.89}]"} {"utterance_id": "1188-133604-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0022.flac", "task_type": "lexical_detection", "query": "fight", "prompt": "When the utterance is complete, report all instances of the word \"fight\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.57, 2.9, "fight"]], "response_mode": "offline", "answer": "[{\"word\":\"fight\",\"start\":2.57,\"end\":2.9}]"} {"utterance_id": "4970-29093-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0017.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.75, 2.0, "six"]], "response_mode": "offline", "answer": "[{\"word\":\"six\",\"start\":1.75,\"end\":2.0}]"} {"utterance_id": "1995-1837-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0026.flac", "task_type": "lexical_counting", "query": "within", "prompt": "When the utterance is complete, report how the word \"within\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.28, 1.6, "one"], [2.7, 2.97, "two"], [5.54, 5.84, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.28,\"end\":1.6},{\"word\":\"two\",\"start\":2.7,\"end\":2.97},{\"word\":\"three\",\"start\":5.54,\"end\":5.84}]"} {"utterance_id": "1580-141083-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0027.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.55, 3.81, "quarter"]], "response_mode": "offline", "answer": "[{\"word\":\"quarter\",\"start\":3.55,\"end\":3.81}]"} {"utterance_id": "5142-33396-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0042.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.8, 4.93, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.8,\"end\":4.93}]"} {"utterance_id": "121-127105-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0032.flac", "task_type": "lexical_detection", "query": "passion", "prompt": "After listening to the whole recording, output every instance of the word \"passion\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.37, 2.96, "passion"]], "response_mode": "offline", "answer": "[{\"word\":\"passion\",\"start\":2.37,\"end\":2.96}]"} {"utterance_id": "7021-85628-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.67, 1.17, "anders"]], "response_mode": "offline", "answer": "[{\"word\":\"anders\",\"start\":0.67,\"end\":1.17}]"} {"utterance_id": "5639-40744-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0038.flac", "task_type": "lexical_detection", "query": "hearts", "prompt": "Process the full audio offline. Find every instance of the word \"hearts\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.84, 13.19, "hearts"]], "response_mode": "offline", "answer": "[{\"word\":\"hearts\",\"start\":12.84,\"end\":13.19}]"} {"utterance_id": "1089-134686-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0037.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Process the full audio offline. Find every instance of a time of day and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.14, 3.63, "dusk"]], "response_mode": "offline", "answer": "[{\"word\":\"dusk\",\"start\":3.14,\"end\":3.63}]"} {"utterance_id": "1089-134686-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0032.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a state or province name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.22, 3.79, "indies"]], "response_mode": "offline", "answer": "[{\"word\":\"indies\",\"start\":3.22,\"end\":3.79}]"} {"utterance_id": "5639-40744-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0037.flac", "task_type": "lexical_detection", "query": "chosen", "prompt": "When the utterance is complete, report all instances of the word \"chosen\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.12, 8.55, "chosen"]], "response_mode": "offline", "answer": "[{\"word\":\"chosen\",\"start\":8.12,\"end\":8.55}]"} {"utterance_id": "4446-2275-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0036.flac", "task_type": "lexical_detection", "query": "roughly", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"roughly\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.67, 1.02, "roughly"]], "response_mode": "offline", "answer": "[{\"word\":\"roughly\",\"start\":0.67,\"end\":1.02}]"} {"utterance_id": "260-123286-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0023.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.53, 4.85, "twenty"]], "response_mode": "offline", "answer": "[{\"word\":\"twenty\",\"start\":4.53,\"end\":4.85}]"} {"utterance_id": "260-123288-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0021.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.57, 2.19, "heads"]], "response_mode": "offline", "answer": "[{\"word\":\"heads\",\"start\":1.57,\"end\":2.19}]"} {"utterance_id": "1320-122617-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0028.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.22, 6.61, "david"]], "response_mode": "offline", "answer": "[{\"word\":\"david\",\"start\":6.22,\"end\":6.61}]"} {"utterance_id": "4970-29093-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0023.flac", "task_type": "lexical_detection", "query": "frontier", "prompt": "Process the full audio offline. Find every instance of the word \"frontier\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.04, 2.82, "frontier"]], "response_mode": "offline", "answer": "[{\"word\":\"frontier\",\"start\":2.04,\"end\":2.82}]"} {"utterance_id": "7176-88083-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0011.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "When the utterance is complete, report all instances of an animal name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.23, 3.58, "trout's"], [7.87, 8.29, "hawk"]], "response_mode": "offline", "answer": "[{\"word\":\"trout's\",\"start\":3.23,\"end\":3.58},{\"word\":\"hawk\",\"start\":7.87,\"end\":8.29}]"} {"utterance_id": "6829-68771-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0031.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 0.71, "eyes"], [1.81, 2.23, "hands"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":0.45,\"end\":0.71},{\"word\":\"hands\",\"start\":1.81,\"end\":2.23}]"} {"utterance_id": "1580-141083-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0027.flac", "task_type": "lexical_detection", "query": "using", "prompt": "When the utterance is complete, report all instances of the word \"using\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.75, 2.03, "using"]], "response_mode": "offline", "answer": "[{\"word\":\"using\",\"start\":1.75,\"end\":2.03}]"} {"utterance_id": "6829-68771-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0000.flac", "task_type": "lexical_detection", "query": "democratic", "prompt": "When the utterance is complete, report all instances of the word \"democratic\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.86, 2.46, "democratic"]], "response_mode": "offline", "answer": "[{\"word\":\"democratic\",\"start\":1.86,\"end\":2.46}]"} {"utterance_id": "7729-102255-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0006.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "After listening to the whole recording, output every instance of a state or province name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[15.92, 16.23, "state"]], "response_mode": "offline", "answer": "[{\"word\":\"state\",\"start\":15.92,\"end\":16.23}]"} {"utterance_id": "8224-274384-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0005.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.55, 4.01, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":3.55,\"end\":4.01}]"} {"utterance_id": "1995-1837-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0023.flac", "task_type": "lexical_counting", "query": "creeping", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"creeping\" with word/start/end. If none, reply [].", "labels": [[3.34, 3.76, "one"], [11.5, 11.93, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.34,\"end\":3.76},{\"word\":\"two\",\"start\":11.5,\"end\":11.93}]"} {"utterance_id": "908-157963-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0021.flac", "task_type": "lexical_detection", "query": "worm", "prompt": "Process the full audio offline. Find every instance of the word \"worm\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.13, 6.56, "worm"]], "response_mode": "offline", "answer": "[{\"word\":\"worm\",\"start\":6.13,\"end\":6.56}]"} {"utterance_id": "5142-36600-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36600/5142-36600-0001.flac", "task_type": "lexical_counting", "query": "whether", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"whether\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.83, 1.08, "one"], [11.86, 12.11, "two"], [15.52, 15.76, "three"], [18.58, 18.88, "four"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.83,\"end\":1.08},{\"word\":\"two\",\"start\":11.86,\"end\":12.11},{\"word\":\"three\",\"start\":15.52,\"end\":15.76},{\"word\":\"four\",\"start\":18.58,\"end\":18.88}]"} {"utterance_id": "6930-81414-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0010.flac", "task_type": "lexical_detection", "query": "voices", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"voices\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.23, 2.03, "voices"]], "response_mode": "offline", "answer": "[{\"word\":\"voices\",\"start\":1.23,\"end\":2.03}]"} {"utterance_id": "2300-131720-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0040.flac", "task_type": "lexical_detection", "query": "technical", "prompt": "After listening to the whole recording, output every instance of the word \"technical\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.62, 2.13, "technical"]], "response_mode": "offline", "answer": "[{\"word\":\"technical\",\"start\":1.62,\"end\":2.13}]"} {"utterance_id": "61-70968-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0014.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.86, 5.24, "twice"]], "response_mode": "offline", "answer": "[{\"word\":\"twice\",\"start\":4.86,\"end\":5.24}]"} {"utterance_id": "8463-294828-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0032.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.8, 2.41, "professor"]], "response_mode": "offline", "answer": "[{\"word\":\"professor\",\"start\":1.8,\"end\":2.41}]"} {"utterance_id": "3575-170457-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0041.flac", "task_type": "semantic_detection", "query": "city", "prompt": "When the utterance is complete, report all instances of a city name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.66, 2.05, "village"]], "response_mode": "offline", "answer": "[{\"word\":\"village\",\"start\":1.66,\"end\":2.05}]"} {"utterance_id": "1995-1836-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0014.flac", "task_type": "lexical_detection", "query": "vanderpool", "prompt": "Process the full audio offline. Find every instance of the word \"vanderpool\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.81, 2.57, "vanderpool"]], "response_mode": "offline", "answer": "[{\"word\":\"vanderpool\",\"start\":1.81,\"end\":2.57}]"} {"utterance_id": "61-70968-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0033.flac", "task_type": "lexical_detection", "query": "entrance", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"entrance\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.43, 1.77, "entrance"]], "response_mode": "offline", "answer": "[{\"word\":\"entrance\",\"start\":1.43,\"end\":1.77}]"} {"utterance_id": "1221-135767-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0009.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a body part: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[4.69, 5.07, "one"], [6.12, 6.52, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.69,\"end\":5.07},{\"word\":\"two\",\"start\":6.12,\"end\":6.52}]"} {"utterance_id": "5683-32879-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0006.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.11, 5.58, "rachel"]], "response_mode": "offline", "answer": "[{\"word\":\"rachel\",\"start\":5.11,\"end\":5.58}]"} {"utterance_id": "4077-13754-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0005.flac", "task_type": "lexical_detection", "query": "constitution", "prompt": "Process the full audio offline. Find every instance of the word \"constitution\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.88, 5.73, "constitution"]], "response_mode": "offline", "answer": "[{\"word\":\"constitution\",\"start\":4.88,\"end\":5.73}]"} {"utterance_id": "4446-2273-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0013.flac", "task_type": "lexical_detection", "query": "late", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"late\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.58, 1.76, "late"]], "response_mode": "offline", "answer": "[{\"word\":\"late\",\"start\":1.58,\"end\":1.76}]"} {"utterance_id": "7729-102255-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0032.flac", "task_type": "lexical_detection", "query": "seventy", "prompt": "When the utterance is complete, report all instances of the word \"seventy\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.57, 5.06, "seventy"]], "response_mode": "offline", "answer": "[{\"word\":\"seventy\",\"start\":4.57,\"end\":5.06}]"} {"utterance_id": "908-157963-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0002.flac", "task_type": "lexical_detection", "query": "water", "prompt": "After listening to the whole recording, output every instance of the word \"water\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.85, 2.41, "water"]], "response_mode": "offline", "answer": "[{\"word\":\"water\",\"start\":1.85,\"end\":2.41}]"} {"utterance_id": "260-123440-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0007.flac", "task_type": "lexical_detection", "query": "feeling", "prompt": "After the audio ends, extract all occurrences of the word \"feeling\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.77, 2.22, "feeling"]], "response_mode": "offline", "answer": "[{\"word\":\"feeling\",\"start\":1.77,\"end\":2.22}]"} {"utterance_id": "8230-279154-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0031.flac", "task_type": "lexical_detection", "query": "occurrences", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"occurrences\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.69, 3.19, "occurrences"], [8.7, 9.61, "occurrences"]], "response_mode": "offline", "answer": "[{\"word\":\"occurrences\",\"start\":2.69,\"end\":3.19},{\"word\":\"occurrences\",\"start\":8.7,\"end\":9.61}]"} {"utterance_id": "1284-1180-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0026.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.66, 4.2, "servant"]], "response_mode": "offline", "answer": "[{\"word\":\"servant\",\"start\":3.66,\"end\":4.2}]"} {"utterance_id": "2830-3980-0061", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0061.flac", "task_type": "lexical_detection", "query": "himself", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"himself\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.64, 5.35, "himself"]], "response_mode": "offline", "answer": "[{\"word\":\"himself\",\"start\":4.64,\"end\":5.35}]"} {"utterance_id": "8455-210777-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0002.flac", "task_type": "lexical_detection", "query": "arriving", "prompt": "After the audio ends, extract all occurrences of the word \"arriving\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.69, 1.11, "arriving"]], "response_mode": "offline", "answer": "[{\"word\":\"arriving\",\"start\":0.69,\"end\":1.11}]"} {"utterance_id": "7729-102255-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0013.flac", "task_type": "lexical_detection", "query": "fact", "prompt": "Process the full audio offline. Find every instance of the word \"fact\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.73, 1.05, "fact"]], "response_mode": "offline", "answer": "[{\"word\":\"fact\",\"start\":0.73,\"end\":1.05}]"} {"utterance_id": "8224-274384-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0001.flac", "task_type": "lexical_detection", "query": "scottish", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"scottish\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.5, 1.09, "scottish"]], "response_mode": "offline", "answer": "[{\"word\":\"scottish\",\"start\":0.5,\"end\":1.09}]"} {"utterance_id": "7729-102255-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0027.flac", "task_type": "semantic_detection", "query": "date", "prompt": "After listening to the whole recording, output every instance of a date with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.33, 5.64, "may twelfth eighteen fifty six"]], "response_mode": "offline", "answer": "[{\"word\":\"may twelfth eighteen fifty six\",\"start\":3.33,\"end\":5.64}]"} {"utterance_id": "237-126133-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0008.flac", "task_type": "lexical_detection", "query": "polly's", "prompt": "Process the full audio offline. Find every instance of the word \"polly's\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.57, 3.05, "polly's"]], "response_mode": "offline", "answer": "[{\"word\":\"polly's\",\"start\":2.57,\"end\":3.05}]"} {"utterance_id": "6829-68769-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0044.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.01, 1.67, "twice sixty"]], "response_mode": "offline", "answer": "[{\"word\":\"twice sixty\",\"start\":1.01,\"end\":1.67}]"} {"utterance_id": "2830-3980-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0011.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.62, 2.18, "apostles"]], "response_mode": "offline", "answer": "[{\"word\":\"apostles\",\"start\":1.62,\"end\":2.18}]"} {"utterance_id": "237-134500-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0010.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Process the full audio offline. Find every instance of a plant name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.16, 3.52, "grass"]], "response_mode": "offline", "answer": "[{\"word\":\"grass\",\"start\":3.16,\"end\":3.52}]"} {"utterance_id": "7176-88083-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0020.flac", "task_type": "lexical_detection", "query": "instant", "prompt": "When the utterance is complete, report all instances of the word \"instant\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.1, 1.59, "instant"]], "response_mode": "offline", "answer": "[{\"word\":\"instant\",\"start\":1.1,\"end\":1.59}]"} {"utterance_id": "7176-92135-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0007.flac", "task_type": "lexical_detection", "query": "dialogue", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"dialogue\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.27, 2.74, "dialogue"]], "response_mode": "offline", "answer": "[{\"word\":\"dialogue\",\"start\":2.27,\"end\":2.74}]"} {"utterance_id": "8230-279154-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0010.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.28, 11.77, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":11.28,\"end\":11.77}]"} {"utterance_id": "4507-16021-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0017.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "Listen to the entire clip, then report the count sequence for a profession as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.58, 2.52, "one"], [6.02, 6.76, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.58,\"end\":2.52},{\"word\":\"two\",\"start\":6.02,\"end\":6.76}]"} {"utterance_id": "1221-135766-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0013.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.26, 0.66, "pearl"]], "response_mode": "offline", "answer": "[{\"word\":\"pearl\",\"start\":0.26,\"end\":0.66}]"} {"utterance_id": "4446-2275-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0043.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a body part with word/start/end. If none, reply [].", "labels": [[2.12, 2.52, "one"], [3.54, 3.92, "two"], [5.0, 5.47, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.12,\"end\":2.52},{\"word\":\"two\",\"start\":3.54,\"end\":3.92},{\"word\":\"three\",\"start\":5.0,\"end\":5.47}]"} {"utterance_id": "6829-68769-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0006.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen to the full audio. After it finishes, find every instance of a money or currency term. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.47, 4.83, "check"]], "response_mode": "offline", "answer": "[{\"word\":\"check\",\"start\":4.47,\"end\":4.83}]"} {"utterance_id": "6829-68771-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0026.flac", "task_type": "lexical_detection", "query": "eliza", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"eliza\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.36, 0.74, "eliza"]], "response_mode": "offline", "answer": "[{\"word\":\"eliza\",\"start\":0.36,\"end\":0.74}]"} {"utterance_id": "6829-68771-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0001.flac", "task_type": "lexical_detection", "query": "accomplished", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"accomplished\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.16, 9.89, "accomplished"]], "response_mode": "offline", "answer": "[{\"word\":\"accomplished\",\"start\":9.16,\"end\":9.89}]"} {"utterance_id": "5639-40744-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0001.flac", "task_type": "semantic_detection", "query": "city", "prompt": "You will hear a complete utterance. When it ends, list every instance of a city name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.23, 4.58, "city"]], "response_mode": "offline", "answer": "[{\"word\":\"city\",\"start\":4.23,\"end\":4.58}]"} {"utterance_id": "5142-36377-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.67, 4.14, "naomi"]], "response_mode": "offline", "answer": "[{\"word\":\"naomi\",\"start\":3.67,\"end\":4.14}]"} {"utterance_id": "2094-142345-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0015.flac", "task_type": "lexical_detection", "query": "molly", "prompt": "Once the audio has finished, identify every instance of the word \"molly\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.68, 2.08, "molly"]], "response_mode": "offline", "answer": "[{\"word\":\"molly\",\"start\":1.68,\"end\":2.08}]"} {"utterance_id": "1089-134691-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0022.flac", "task_type": "lexical_detection", "query": "deference", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"deference\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.55, 1.99, "deference"]], "response_mode": "offline", "answer": "[{\"word\":\"deference\",\"start\":1.55,\"end\":1.99}]"} {"utterance_id": "260-123440-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0019.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen to the full audio. After it finishes, find every instance of an animal name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.79, 2.07, "mouse"]], "response_mode": "offline", "answer": "[{\"word\":\"mouse\",\"start\":1.79,\"end\":2.07}]"} {"utterance_id": "8463-294828-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0031.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.33, 0.5, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.33,\"end\":0.5}]"} {"utterance_id": "2830-3980-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0053.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.62, 1.86, "paul"], [3.35, 3.78, "christ"]], "response_mode": "offline", "answer": "[{\"word\":\"paul\",\"start\":1.62,\"end\":1.86},{\"word\":\"christ\",\"start\":3.35,\"end\":3.78}]"} {"utterance_id": "1320-122612-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0006.flac", "task_type": "lexical_detection", "query": "examine", "prompt": "When the utterance is complete, report all instances of the word \"examine\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.02, 2.56, "examine"]], "response_mode": "offline", "answer": "[{\"word\":\"examine\",\"start\":2.02,\"end\":2.56}]"} {"utterance_id": "1580-141084-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0005.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.77, 2.28, "holmes"]], "response_mode": "offline", "answer": "[{\"word\":\"holmes\",\"start\":1.77,\"end\":2.28}]"} {"utterance_id": "2300-131720-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0002.flac", "task_type": "lexical_detection", "query": "data", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"data\" with fields word, start, end (seconds). If none, reply [].", "labels": [[8.59, 9.06, "data"]], "response_mode": "offline", "answer": "[{\"word\":\"data\",\"start\":8.59,\"end\":9.06}]"} {"utterance_id": "1995-1826-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0013.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an ordinal number with fields word, start, end (seconds). If none, reply [].", "labels": [[0.7, 1.1, "hundredth"]], "response_mode": "offline", "answer": "[{\"word\":\"hundredth\",\"start\":0.7,\"end\":1.1}]"} {"utterance_id": "4446-2273-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.22, 6.54, "mike"]], "response_mode": "offline", "answer": "[{\"word\":\"mike\",\"start\":6.22,\"end\":6.54}]"} {"utterance_id": "2300-131720-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0012.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.35, 0.86, "edison"]], "response_mode": "offline", "answer": "[{\"word\":\"edison\",\"start\":0.35,\"end\":0.86}]"} {"utterance_id": "1995-1837-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0015.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "You will hear a complete utterance. When it ends, list every instance of a plant name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.75, 1.16, "cotton"], [3.71, 4.36, "bolls"]], "response_mode": "offline", "answer": "[{\"word\":\"cotton\",\"start\":0.75,\"end\":1.16},{\"word\":\"bolls\",\"start\":3.71,\"end\":4.36}]"} {"utterance_id": "260-123288-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0019.flac", "task_type": "lexical_detection", "query": "violence", "prompt": "After the audio ends, extract all occurrences of the word \"violence\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.11, 1.54, "violence"]], "response_mode": "offline", "answer": "[{\"word\":\"violence\",\"start\":1.11,\"end\":1.54}]"} {"utterance_id": "1089-134686-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0013.flac", "task_type": "lexical_detection", "query": "impelled", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"impelled\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.34, 1.86, "impelled"]], "response_mode": "offline", "answer": "[{\"word\":\"impelled\",\"start\":1.34,\"end\":1.86}]"} {"utterance_id": "7021-85628-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0027.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After the audio ends, extract all occurrences of a color name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.58, 7.85, "red"]], "response_mode": "offline", "answer": "[{\"word\":\"red\",\"start\":7.58,\"end\":7.85}]"} {"utterance_id": "4507-16021-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0038.flac", "task_type": "lexical_detection", "query": "becomes", "prompt": "After listening to the whole recording, output every instance of the word \"becomes\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.59, 2.27, "becomes"]], "response_mode": "offline", "answer": "[{\"word\":\"becomes\",\"start\":1.59,\"end\":2.27}]"} {"utterance_id": "7176-88083-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0006.flac", "task_type": "semantic_detection", "query": "food", "prompt": "You will hear a complete utterance. When it ends, list every instance of a food name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.71, 4.07, "meal"]], "response_mode": "offline", "answer": "[{\"word\":\"meal\",\"start\":3.71,\"end\":4.07}]"} {"utterance_id": "3729-6852-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0009.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[5.06, 5.33, "ten"]], "response_mode": "offline", "answer": "[{\"word\":\"ten\",\"start\":5.06,\"end\":5.33}]"} {"utterance_id": "1221-135767-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0024.flac", "task_type": "lexical_detection", "query": "bushes", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"bushes\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.66, 2.1, "bushes"]], "response_mode": "offline", "answer": "[{\"word\":\"bushes\",\"start\":1.66,\"end\":2.1}]"} {"utterance_id": "4446-2271-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0016.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "After the audio ends, extract all occurrences of an animal name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.46, 4.8, "donkey"]], "response_mode": "offline", "answer": "[{\"word\":\"donkey\",\"start\":4.46,\"end\":4.8}]"} {"utterance_id": "2300-131720-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0022.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.67, 3.9, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":3.67,\"end\":3.9}]"} {"utterance_id": "260-123288-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0003.flac", "task_type": "lexical_detection", "query": "penetrate", "prompt": "Once the audio has finished, identify every instance of the word \"penetrate\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.16, 2.74, "penetrate"]], "response_mode": "offline", "answer": "[{\"word\":\"penetrate\",\"start\":2.16,\"end\":2.74}]"} {"utterance_id": "3729-6852-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0046.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.56, 1.2, "housekeeper"]], "response_mode": "offline", "answer": "[{\"word\":\"housekeeper\",\"start\":0.56,\"end\":1.2}]"} {"utterance_id": "4970-29095-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0037.flac", "task_type": "lexical_detection", "query": "wise", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"wise\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.89, 1.33, "wise"]], "response_mode": "offline", "answer": "[{\"word\":\"wise\",\"start\":0.89,\"end\":1.33}]"} {"utterance_id": "1995-1826-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0021.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.68, 6.22, "cotton"]], "response_mode": "offline", "answer": "[{\"word\":\"cotton\",\"start\":5.68,\"end\":6.22}]"} {"utterance_id": "5142-36377-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0024.flac", "task_type": "lexical_detection", "query": "children", "prompt": "Once the audio has finished, identify every instance of the word \"children\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.1, 5.7, "children"]], "response_mode": "offline", "answer": "[{\"word\":\"children\",\"start\":5.1,\"end\":5.7}]"} {"utterance_id": "7021-85628-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0004.flac", "task_type": "lexical_detection", "query": "anders", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"anders\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.99, 2.55, "anders"]], "response_mode": "offline", "answer": "[{\"word\":\"anders\",\"start\":1.99,\"end\":2.55}]"} {"utterance_id": "8555-292519-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0009.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "After the audio ends, extract all occurrences of a weather term. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.22, 13.98, "wind gusts"]], "response_mode": "offline", "answer": "[{\"word\":\"wind gusts\",\"start\":13.22,\"end\":13.98}]"} {"utterance_id": "7127-75947-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0002.flac", "task_type": "lexical_detection", "query": "indifference", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"indifference\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.23, 2.88, "indifference"]], "response_mode": "offline", "answer": "[{\"word\":\"indifference\",\"start\":2.23,\"end\":2.88}]"} {"utterance_id": "908-31957-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0006.flac", "task_type": "lexical_counting", "query": "thy", "prompt": "Listen to the entire clip, then report the count sequence for the word \"thy\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.73, 0.88, "one"], [4.71, 4.92, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.73,\"end\":0.88},{\"word\":\"two\",\"start\":4.71,\"end\":4.92}]"} {"utterance_id": "4507-16021-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0053.flac", "task_type": "lexical_detection", "query": "shady", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"shady\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.25, 2.95, "shady"]], "response_mode": "offline", "answer": "[{\"word\":\"shady\",\"start\":2.25,\"end\":2.95}]"} {"utterance_id": "237-134500-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0031.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.98, 4.29, "four"]], "response_mode": "offline", "answer": "[{\"word\":\"four\",\"start\":3.98,\"end\":4.29}]"} {"utterance_id": "5683-32866-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0028.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Process the full audio offline. Find every instance of a color name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.12, 4.57, "black"]], "response_mode": "offline", "answer": "[{\"word\":\"black\",\"start\":4.12,\"end\":4.57}]"} {"utterance_id": "4446-2273-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0016.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.6, 2.19, "flax"]], "response_mode": "offline", "answer": "[{\"word\":\"flax\",\"start\":1.6,\"end\":2.19}]"} {"utterance_id": "1995-1836-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.57, 3.35, "mary taylor"]], "response_mode": "offline", "answer": "[{\"word\":\"mary taylor\",\"start\":2.57,\"end\":3.35}]"} {"utterance_id": "3570-5694-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0003.flac", "task_type": "lexical_counting", "query": "tabu", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"tabu\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[2.38, 2.69, "one"], [12.02, 12.55, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.38,\"end\":2.69},{\"word\":\"two\",\"start\":12.02,\"end\":12.55}]"} {"utterance_id": "8224-274381-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0014.flac", "task_type": "semantic_counting", "query": "organization", "prompt": "Listen to the full audio. Count every occurrence of an organization name. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.99, 2.57, "one"], [13.97, 14.44, "two"], [16.75, 17.39, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.99,\"end\":2.57},{\"word\":\"two\",\"start\":13.97,\"end\":14.44},{\"word\":\"three\",\"start\":16.75,\"end\":17.39}]"} {"utterance_id": "61-70968-0058", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0058.flac", "task_type": "lexical_detection", "query": "forgive", "prompt": "When the utterance is complete, report all instances of the word \"forgive\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.68, 1.15, "forgive"]], "response_mode": "offline", "answer": "[{\"word\":\"forgive\",\"start\":0.68,\"end\":1.15}]"} {"utterance_id": "4446-2275-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0037.flac", "task_type": "lexical_detection", "query": "bartley", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"bartley\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.38, 0.87, "bartley"]], "response_mode": "offline", "answer": "[{\"word\":\"bartley\",\"start\":0.38,\"end\":0.87}]"} {"utterance_id": "4970-29095-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0010.flac", "task_type": "lexical_detection", "query": "medicine", "prompt": "After listening to the whole recording, output every instance of the word \"medicine\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.57, 2.41, "medicine"]], "response_mode": "offline", "answer": "[{\"word\":\"medicine\",\"start\":1.57,\"end\":2.41}]"} {"utterance_id": "260-123286-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0018.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.02, 3.17, "one"], [4.29, 4.55, "thirty"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.02,\"end\":3.17},{\"word\":\"thirty\",\"start\":4.29,\"end\":4.55}]"} {"utterance_id": "7021-79730-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0004.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Process the full audio offline. For every occurrence of a body part, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.22, 7.96, "one"], [13.44, 13.95, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.22,\"end\":7.96},{\"word\":\"two\",\"start\":13.44,\"end\":13.95}]"} {"utterance_id": "8230-279154-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0043.flac", "task_type": "lexical_detection", "query": "primitive", "prompt": "When the utterance is complete, report all instances of the word \"primitive\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[17.13, 17.82, "primitive"]], "response_mode": "offline", "answer": "[{\"word\":\"primitive\",\"start\":17.13,\"end\":17.82}]"} {"utterance_id": "7127-75947-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0009.flac", "task_type": "lexical_detection", "query": "excitement", "prompt": "After listening to the whole recording, output every instance of the word \"excitement\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.87, 10.5, "excitement"]], "response_mode": "offline", "answer": "[{\"word\":\"excitement\",\"start\":9.87,\"end\":10.5}]"} {"utterance_id": "8230-279154-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0041.flac", "task_type": "lexical_detection", "query": "definition", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"definition\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.66, 1.29, "definition"]], "response_mode": "offline", "answer": "[{\"word\":\"definition\",\"start\":0.66,\"end\":1.29}]"} {"utterance_id": "1995-1837-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0012.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.21, 7.48, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":7.21,\"end\":7.48}]"} {"utterance_id": "2830-3979-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0001.flac", "task_type": "lexical_detection", "query": "permitted", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"permitted\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.73, 2.21, "permitted"]], "response_mode": "offline", "answer": "[{\"word\":\"permitted\",\"start\":1.73,\"end\":2.21}]"} {"utterance_id": "8455-210777-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0012.flac", "task_type": "lexical_detection", "query": "proud", "prompt": "Once the audio has finished, identify every instance of the word \"proud\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.19, 2.48, "proud"]], "response_mode": "offline", "answer": "[{\"word\":\"proud\",\"start\":2.19,\"end\":2.48}]"} {"utterance_id": "6930-81414-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0018.flac", "task_type": "lexical_detection", "query": "saying", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"saying\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.66, 1.13, "saying"]], "response_mode": "offline", "answer": "[{\"word\":\"saying\",\"start\":0.66,\"end\":1.13}]"} {"utterance_id": "4446-2275-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0025.flac", "task_type": "lexical_detection", "query": "happy", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"happy\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.67, 0.98, "happy"]], "response_mode": "offline", "answer": "[{\"word\":\"happy\",\"start\":0.67,\"end\":0.98}]"} {"utterance_id": "5639-40744-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0005.flac", "task_type": "lexical_detection", "query": "woe", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"woe\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.87, 5.32, "woe"]], "response_mode": "offline", "answer": "[{\"word\":\"woe\",\"start\":4.87,\"end\":5.32}]"} {"utterance_id": "6930-81414-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0014.flac", "task_type": "semantic_detection", "query": "color", "prompt": "You will hear a complete utterance. When it ends, list every instance of a color name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.93, 3.42, "red"]], "response_mode": "offline", "answer": "[{\"word\":\"red\",\"start\":2.93,\"end\":3.42}]"} {"utterance_id": "1284-1181-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0010.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the full audio. After it finishes, find every instance of a color name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.37, 5.64, "white"]], "response_mode": "offline", "answer": "[{\"word\":\"white\",\"start\":5.37,\"end\":5.64}]"} {"utterance_id": "4992-41797-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.54, 1.37, "mister popham"]], "response_mode": "offline", "answer": "[{\"word\":\"mister popham\",\"start\":0.54,\"end\":1.37}]"} {"utterance_id": "5142-36377-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0025.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.39, 0.95, "silas"], [2.94, 3.36, "ambrose"], [5.91, 6.34, "naomi"]], "response_mode": "offline", "answer": "[{\"word\":\"silas\",\"start\":0.39,\"end\":0.95},{\"word\":\"ambrose\",\"start\":2.94,\"end\":3.36},{\"word\":\"naomi\",\"start\":5.91,\"end\":6.34}]"} {"utterance_id": "4446-2273-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0010.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "After the audio ends, tally every instance of an animal name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[1.75, 2.35, "one"], [6.5, 7.19, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.75,\"end\":2.35},{\"word\":\"two\",\"start\":6.5,\"end\":7.19}]"} {"utterance_id": "7176-92135-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0023.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.57, 5.58, "horatio hamlet"]], "response_mode": "offline", "answer": "[{\"word\":\"horatio hamlet\",\"start\":4.57,\"end\":5.58}]"} {"utterance_id": "4077-13754-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0016.flac", "task_type": "lexical_counting", "query": "new", "prompt": "After listening to the whole recording, output timed count words for the word \"new\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[7.87, 8.09, "one"], [9.99, 10.14, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.87,\"end\":8.09},{\"word\":\"two\",\"start\":9.99,\"end\":10.14}]"} {"utterance_id": "237-126133-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0012.flac", "task_type": "lexical_detection", "query": "brown", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"brown\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.07, 3.53, "brown"]], "response_mode": "offline", "answer": "[{\"word\":\"brown\",\"start\":3.07,\"end\":3.53}]"} {"utterance_id": "7021-79730-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0003.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.79, 15.31, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":14.79,\"end\":15.31}]"} {"utterance_id": "4970-29093-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0005.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Do not answer until the audio ends. Then list each instance of an animal name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[16.76, 17.42, "nightingale's"], [19.45, 19.93, "bul bul"]], "response_mode": "offline", "answer": "[{\"word\":\"nightingale's\",\"start\":16.76,\"end\":17.42},{\"word\":\"bul bul\",\"start\":19.45,\"end\":19.93}]"} {"utterance_id": "1995-1826-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0019.flac", "task_type": "lexical_detection", "query": "boys", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"boys\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.91, 3.5, "boys"]], "response_mode": "offline", "answer": "[{\"word\":\"boys\",\"start\":2.91,\"end\":3.5}]"} {"utterance_id": "4446-2273-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0010.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After the audio ends, extract all occurrences of a color name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.99, 9.33, "yellow"]], "response_mode": "offline", "answer": "[{\"word\":\"yellow\",\"start\":8.99,\"end\":9.33}]"} {"utterance_id": "3570-5696-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0009.flac", "task_type": "semantic_counting", "query": "money", "prompt": "When the utterance is complete, report how a money or currency term accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.65, 5.31, "one"], [7.19, 7.78, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.65,\"end\":5.31},{\"word\":\"two\",\"start\":7.19,\"end\":7.78}]"} {"utterance_id": "6829-68769-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0031.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After listening to the whole recording, output every instance of an organization name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.16, 2.75, "company"]], "response_mode": "offline", "answer": "[{\"word\":\"company\",\"start\":2.16,\"end\":2.75}]"} {"utterance_id": "5105-28240-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0017.flac", "task_type": "lexical_detection", "query": "mysterious", "prompt": "Process the full audio offline. Find every instance of the word \"mysterious\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.82, 1.56, "mysterious"]], "response_mode": "offline", "answer": "[{\"word\":\"mysterious\",\"start\":0.82,\"end\":1.56}]"} {"utterance_id": "4077-13754-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0011.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Listen to the full audio. After it finishes, find every instance of a state or province name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.68, 3.12, "utah"]], "response_mode": "offline", "answer": "[{\"word\":\"utah\",\"start\":2.68,\"end\":3.12}]"} {"utterance_id": "5683-32865-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0014.flac", "task_type": "lexical_detection", "query": "quarters", "prompt": "Once the audio has finished, identify every instance of the word \"quarters\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.88, 2.48, "quarters"]], "response_mode": "offline", "answer": "[{\"word\":\"quarters\",\"start\":1.88,\"end\":2.48}]"} {"utterance_id": "61-70968-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0051.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Once the audio has finished, identify every instance of a family relation. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.91, 2.44, "child"]], "response_mode": "offline", "answer": "[{\"word\":\"child\",\"start\":1.91,\"end\":2.44}]"} {"utterance_id": "3575-170457-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0008.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.04, 6.33, "twenty"]], "response_mode": "offline", "answer": "[{\"word\":\"twenty\",\"start\":6.04,\"end\":6.33}]"} {"utterance_id": "4077-13751-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0019.flac", "task_type": "lexical_detection", "query": "mormons", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"mormons\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.08, 2.65, "mormons"]], "response_mode": "offline", "answer": "[{\"word\":\"mormons\",\"start\":2.08,\"end\":2.65}]"} {"utterance_id": "8555-284449-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0010.flac", "task_type": "lexical_detection", "query": "satisfaction", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"satisfaction\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.21, 2.25, "satisfaction"]], "response_mode": "offline", "answer": "[{\"word\":\"satisfaction\",\"start\":1.21,\"end\":2.25}]"} {"utterance_id": "7176-88083-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0005.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.04, 1.31, "wing"]], "response_mode": "offline", "answer": "[{\"word\":\"wing\",\"start\":1.04,\"end\":1.31}]"} {"utterance_id": "1284-1181-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0003.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.36, 0.67, "hair"], [2.8, 3.11, "neck"]], "response_mode": "offline", "answer": "[{\"word\":\"hair\",\"start\":0.36,\"end\":0.67},{\"word\":\"neck\",\"start\":2.8,\"end\":3.11}]"} {"utterance_id": "260-123288-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0028.flac", "task_type": "lexical_detection", "query": "pains", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"pains\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.51, 2.32, "pains"]], "response_mode": "offline", "answer": "[{\"word\":\"pains\",\"start\":1.51,\"end\":2.32}]"} {"utterance_id": "121-127105-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0035.flac", "task_type": "lexical_detection", "query": "sacrifice", "prompt": "After the audio ends, extract all occurrences of the word \"sacrifice\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.62, 11.4, "sacrifice"]], "response_mode": "offline", "answer": "[{\"word\":\"sacrifice\",\"start\":10.62,\"end\":11.4}]"} {"utterance_id": "4446-2275-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0029.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.32, 1.8, "bartley"]], "response_mode": "offline", "answer": "[{\"word\":\"bartley\",\"start\":1.32,\"end\":1.8}]"} {"utterance_id": "2830-3979-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.54, 2.97, "luther's"]], "response_mode": "offline", "answer": "[{\"word\":\"luther's\",\"start\":2.54,\"end\":2.97}]"} {"utterance_id": "8463-287645-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.42, 2.24, "john wesley"]], "response_mode": "offline", "answer": "[{\"word\":\"john wesley\",\"start\":1.42,\"end\":2.24}]"} {"utterance_id": "2300-131720-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0003.flac", "task_type": "semantic_detection", "query": "date", "prompt": "Process the full audio offline. Find every instance of a date and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.63, 14.92, "eighteen ninety nine"]], "response_mode": "offline", "answer": "[{\"word\":\"eighteen ninety nine\",\"start\":13.63,\"end\":14.92}]"} {"utterance_id": "61-70968-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0017.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.73, 3.89, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.73,\"end\":3.89}]"} {"utterance_id": "2300-131720-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0028.flac", "task_type": "lexical_detection", "query": "inventors", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"inventors\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.19, 5.76, "inventors"]], "response_mode": "offline", "answer": "[{\"word\":\"inventors\",\"start\":5.19,\"end\":5.76}]"} {"utterance_id": "121-127105-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0022.flac", "task_type": "lexical_detection", "query": "don't", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"don't\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.49, 1.68, "don't"]], "response_mode": "offline", "answer": "[{\"word\":\"don't\",\"start\":1.49,\"end\":1.68}]"} {"utterance_id": "7729-102255-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0045.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.22, 1.12, "captain martin"]], "response_mode": "offline", "answer": "[{\"word\":\"captain martin\",\"start\":0.22,\"end\":1.12}]"} {"utterance_id": "1089-134686-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0031.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the full audio. After it finishes, find every instance of a plant name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.59, 6.11, "flower"]], "response_mode": "offline", "answer": "[{\"word\":\"flower\",\"start\":5.59,\"end\":6.11}]"} {"utterance_id": "3570-5695-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0002.flac", "task_type": "lexical_detection", "query": "scale", "prompt": "When the utterance is complete, report all instances of the word \"scale\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.66, 2.08, "scale"]], "response_mode": "offline", "answer": "[{\"word\":\"scale\",\"start\":1.66,\"end\":2.08}]"} {"utterance_id": "5683-32865-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0003.flac", "task_type": "lexical_detection", "query": "cousins", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"cousins\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.66, 1.22, "cousins"], [2.61, 3.33, "cousins"]], "response_mode": "offline", "answer": "[{\"word\":\"cousins\",\"start\":0.66,\"end\":1.22},{\"word\":\"cousins\",\"start\":2.61,\"end\":3.33}]"} {"utterance_id": "5683-32879-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0014.flac", "task_type": "lexical_detection", "query": "hurriedly", "prompt": "Process the full audio offline. Find every instance of the word \"hurriedly\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.66, 4.25, "hurriedly"]], "response_mode": "offline", "answer": "[{\"word\":\"hurriedly\",\"start\":3.66,\"end\":4.25}]"} {"utterance_id": "8463-294825-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0015.flac", "task_type": "lexical_detection", "query": "eighth", "prompt": "Once the audio has finished, identify every instance of the word \"eighth\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.95, 2.24, "eighth"]], "response_mode": "offline", "answer": "[{\"word\":\"eighth\",\"start\":1.95,\"end\":2.24}]"} {"utterance_id": "2094-142345-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0028.flac", "task_type": "lexical_detection", "query": "iron", "prompt": "When the utterance is complete, report all instances of the word \"iron\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.97, 6.29, "iron"]], "response_mode": "offline", "answer": "[{\"word\":\"iron\",\"start\":5.97,\"end\":6.29}]"} {"utterance_id": "1995-1837-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0020.flac", "task_type": "lexical_detection", "query": "ten", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"ten\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.5, 3.08, "ten"]], "response_mode": "offline", "answer": "[{\"word\":\"ten\",\"start\":2.5,\"end\":3.08}]"} {"utterance_id": "4970-29095-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0031.flac", "task_type": "lexical_detection", "query": "fortunes", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"fortunes\" with fields word, start, end (seconds). If none, reply [].", "labels": [[10.84, 11.29, "fortunes"]], "response_mode": "offline", "answer": "[{\"word\":\"fortunes\",\"start\":10.84,\"end\":11.29}]"} {"utterance_id": "5683-32866-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0022.flac", "task_type": "lexical_detection", "query": "thick", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"thick\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.15, 1.48, "thick"]], "response_mode": "offline", "answer": "[{\"word\":\"thick\",\"start\":1.15,\"end\":1.48}]"} {"utterance_id": "2094-142345-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0013.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.64, 1.07, "tongue"], [2.52, 2.99, "eye"]], "response_mode": "offline", "answer": "[{\"word\":\"tongue\",\"start\":0.64,\"end\":1.07},{\"word\":\"eye\",\"start\":2.52,\"end\":2.99}]"} {"utterance_id": "1320-122612-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0003.flac", "task_type": "lexical_detection", "query": "rivulet", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"rivulet\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.11, 4.58, "rivulet"]], "response_mode": "offline", "answer": "[{\"word\":\"rivulet\",\"start\":4.11,\"end\":4.58}]"} {"utterance_id": "4507-16021-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0048.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.49, 2.98, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":2.49,\"end\":2.98}]"} {"utterance_id": "61-70970-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0037.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.88, 2.35, "warrenton"]], "response_mode": "offline", "answer": "[{\"word\":\"warrenton\",\"start\":1.88,\"end\":2.35}]"} {"utterance_id": "2830-3980-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0013.flac", "task_type": "lexical_detection", "query": "appointed", "prompt": "After the audio ends, extract all occurrences of the word \"appointed\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.76, "appointed"]], "response_mode": "offline", "answer": "[{\"word\":\"appointed\",\"start\":2.35,\"end\":2.76}]"} {"utterance_id": "4992-41806-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0009.flac", "task_type": "lexical_detection", "query": "harmon", "prompt": "Once the audio has finished, identify every instance of the word \"harmon\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.11, 1.47, "harmon"]], "response_mode": "offline", "answer": "[{\"word\":\"harmon\",\"start\":1.11,\"end\":1.47}]"} {"utterance_id": "4077-13751-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0007.flac", "task_type": "lexical_detection", "query": "remnant", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"remnant\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.77, 8.36, "remnant"]], "response_mode": "offline", "answer": "[{\"word\":\"remnant\",\"start\":7.77,\"end\":8.36}]"} {"utterance_id": "3570-5695-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0001.flac", "task_type": "lexical_counting", "query": "duties", "prompt": "Process the full audio offline. For every occurrence of the word \"duties\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.63, 2.99, "one"], [10.16, 10.56, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.63,\"end\":2.99},{\"word\":\"two\",\"start\":10.16,\"end\":10.56}]"} {"utterance_id": "5639-40744-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0026.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 0.89, "luis"]], "response_mode": "offline", "answer": "[{\"word\":\"luis\",\"start\":0.51,\"end\":0.89}]"} {"utterance_id": "2300-131720-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0023.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "When the utterance is complete, report all instances of an ordinal number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.11, 9.57, "fourth"]], "response_mode": "offline", "answer": "[{\"word\":\"fourth\",\"start\":9.11,\"end\":9.57}]"} {"utterance_id": "1995-1836-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0005.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.16, 0.83, "missus grey"]], "response_mode": "offline", "answer": "[{\"word\":\"missus grey\",\"start\":0.16,\"end\":0.83}]"} {"utterance_id": "1089-134691-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[18.91, 19.56, "brother hickey"]], "response_mode": "offline", "answer": "[{\"word\":\"brother hickey\",\"start\":18.91,\"end\":19.56}]"} {"utterance_id": "7021-79730-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0005.flac", "task_type": "lexical_detection", "query": "trouble", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"trouble\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.36, 4.83, "trouble"]], "response_mode": "offline", "answer": "[{\"word\":\"trouble\",\"start\":4.36,\"end\":4.83}]"} {"utterance_id": "1284-1181-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0020.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.98, 3.32, "unc"]], "response_mode": "offline", "answer": "[{\"word\":\"unc\",\"start\":2.98,\"end\":3.32}]"} {"utterance_id": "5639-40744-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0001.flac", "task_type": "lexical_detection", "query": "justice", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"justice\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.45, 3.95, "justice"]], "response_mode": "offline", "answer": "[{\"word\":\"justice\",\"start\":3.45,\"end\":3.95}]"} {"utterance_id": "672-122797-0058", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0058.flac", "task_type": "lexical_detection", "query": "dumpy", "prompt": "After listening to the whole recording, output every instance of the word \"dumpy\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.13, 2.82, "dumpy"]], "response_mode": "offline", "answer": "[{\"word\":\"dumpy\",\"start\":2.13,\"end\":2.82}]"} {"utterance_id": "260-123440-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0018.flac", "task_type": "lexical_detection", "query": "swimming", "prompt": "Once the audio has finished, identify every instance of the word \"swimming\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.1, 1.47, "swimming"]], "response_mode": "offline", "answer": "[{\"word\":\"swimming\",\"start\":1.1,\"end\":1.47}]"} {"utterance_id": "6829-68771-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0033.flac", "task_type": "lexical_detection", "query": "gesture", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"gesture\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.17, 3.66, "gesture"]], "response_mode": "offline", "answer": "[{\"word\":\"gesture\",\"start\":3.17,\"end\":3.66}]"} {"utterance_id": "4446-2273-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0010.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "After the audio ends, extract all occurrences of a state or province name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.33, 9.65, "rhone"]], "response_mode": "offline", "answer": "[{\"word\":\"rhone\",\"start\":9.33,\"end\":9.65}]"} {"utterance_id": "1995-1837-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0016.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Process the full audio offline. Find every instance of a plant name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.36, 6.86, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":6.36,\"end\":6.86}]"} {"utterance_id": "4992-41797-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0015.flac", "task_type": "lexical_detection", "query": "cyril", "prompt": "After the audio ends, extract all occurrences of the word \"cyril\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.21, 0.67, "cyril"]], "response_mode": "offline", "answer": "[{\"word\":\"cyril\",\"start\":0.21,\"end\":0.67}]"} {"utterance_id": "2094-142345-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0042.flac", "task_type": "lexical_detection", "query": "patience", "prompt": "After listening to the whole recording, output every instance of the word \"patience\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.06, 1.51, "patience"]], "response_mode": "offline", "answer": "[{\"word\":\"patience\",\"start\":1.06,\"end\":1.51}]"} {"utterance_id": "6930-76324-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0009.flac", "task_type": "lexical_detection", "query": "copy", "prompt": "When the utterance is complete, report all instances of the word \"copy\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.67, 2.06, "copy"]], "response_mode": "offline", "answer": "[{\"word\":\"copy\",\"start\":1.67,\"end\":2.06}]"} {"utterance_id": "5683-32879-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0002.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.7, 8.1, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":7.7,\"end\":8.1}]"} {"utterance_id": "1580-141083-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0038.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.85, 2.11, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":1.85,\"end\":2.11}]"} {"utterance_id": "1188-133604-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0044.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.42, 8.92, "shepherd"]], "response_mode": "offline", "answer": "[{\"word\":\"shepherd\",\"start\":8.42,\"end\":8.92}]"} {"utterance_id": "2300-131720-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0036.flac", "task_type": "lexical_detection", "query": "ninety", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"ninety\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.35, 3.64, "ninety"]], "response_mode": "offline", "answer": "[{\"word\":\"ninety\",\"start\":3.35,\"end\":3.64}]"} {"utterance_id": "7127-75946-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0025.flac", "task_type": "lexical_detection", "query": "effect", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"effect\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.21, 2.54, "effect"]], "response_mode": "offline", "answer": "[{\"word\":\"effect\",\"start\":2.21,\"end\":2.54}]"} {"utterance_id": "2830-3980-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0044.flac", "task_type": "lexical_detection", "query": "grace", "prompt": "Process the full audio offline. Find every instance of the word \"grace\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.34, 1.75, "grace"]], "response_mode": "offline", "answer": "[{\"word\":\"grace\",\"start\":1.34,\"end\":1.75}]"} {"utterance_id": "8224-274384-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0010.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.49, 11.98, "king"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":11.49,\"end\":11.98}]"} {"utterance_id": "5142-36377-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0013.flac", "task_type": "lexical_detection", "query": "opposite", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"opposite\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.27, 5.74, "opposite"]], "response_mode": "offline", "answer": "[{\"word\":\"opposite\",\"start\":5.27,\"end\":5.74}]"} {"utterance_id": "1995-1826-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0023.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Process the full audio offline. Find every instance of a plant name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.15, 2.7, "vines"], [4.74, 5.0, "roots"]], "response_mode": "offline", "answer": "[{\"word\":\"vines\",\"start\":2.15,\"end\":2.7},{\"word\":\"roots\",\"start\":4.74,\"end\":5.0}]"} {"utterance_id": "5105-28241-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0016.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.32, 1.89, "lieutenant"]], "response_mode": "offline", "answer": "[{\"word\":\"lieutenant\",\"start\":1.32,\"end\":1.89}]"} {"utterance_id": "8555-292519-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0002.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen to the full audio. After it finishes, find every instance of a city name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.48, 1.31, "venice"]], "response_mode": "offline", "answer": "[{\"word\":\"venice\",\"start\":0.48,\"end\":1.31}]"} {"utterance_id": "237-134500-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0034.flac", "task_type": "lexical_detection", "query": "returned", "prompt": "After listening to the whole recording, output every instance of the word \"returned\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.5, 2.01, "returned"]], "response_mode": "offline", "answer": "[{\"word\":\"returned\",\"start\":1.5,\"end\":2.01}]"} {"utterance_id": "6829-68769-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0001.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a person's name with word/start/end. If none, reply [].", "labels": [[1.72, 2.4, "one"], [3.48, 4.15, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.72,\"end\":2.4},{\"word\":\"two\",\"start\":3.48,\"end\":4.15}]"} {"utterance_id": "3575-170457-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0050.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.42, 0.84, "tabby"]], "response_mode": "offline", "answer": "[{\"word\":\"tabby\",\"start\":0.42,\"end\":0.84}]"} {"utterance_id": "1188-133604-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0034.flac", "task_type": "lexical_detection", "query": "complete", "prompt": "Process the full audio offline. Find every instance of the word \"complete\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.55, 9.99, "complete"]], "response_mode": "offline", "answer": "[{\"word\":\"complete\",\"start\":9.55,\"end\":9.99}]"} {"utterance_id": "6829-68769-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0020.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.68, 2.0, "gates"]], "response_mode": "offline", "answer": "[{\"word\":\"gates\",\"start\":1.68,\"end\":2.0}]"} {"utterance_id": "7176-92135-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0001.flac", "task_type": "lexical_detection", "query": "london", "prompt": "After the audio ends, extract all occurrences of the word \"london\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.78, 3.07, "london"]], "response_mode": "offline", "answer": "[{\"word\":\"london\",\"start\":2.78,\"end\":3.07}]"} {"utterance_id": "4446-2273-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0036.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.5, 1.65, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":1.5,\"end\":1.65}]"} {"utterance_id": "7127-75946-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0012.flac", "task_type": "lexical_detection", "query": "news", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"news\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.49, 0.82, "news"]], "response_mode": "offline", "answer": "[{\"word\":\"news\",\"start\":0.49,\"end\":0.82}]"} {"utterance_id": "2094-142345-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0008.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen carefully to the entire clip, then report every instance of an animal name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[16.85, 17.08, "cow"], [25.47, 25.97, "ducks"]], "response_mode": "offline", "answer": "[{\"word\":\"cow\",\"start\":16.85,\"end\":17.08},{\"word\":\"ducks\",\"start\":25.47,\"end\":25.97}]"} {"utterance_id": "2961-961-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0012.flac", "task_type": "semantic_counting", "query": "city", "prompt": "After the audio ends, tally every instance of a city name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[3.51, 3.99, "one"], [5.61, 6.21, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.51,\"end\":3.99},{\"word\":\"two\",\"start\":5.61,\"end\":6.21}]"} {"utterance_id": "1580-141084-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0013.flac", "task_type": "lexical_detection", "query": "worst", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"worst\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.11, 1.49, "worst"]], "response_mode": "offline", "answer": "[{\"word\":\"worst\",\"start\":1.11,\"end\":1.49}]"} {"utterance_id": "4992-41806-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0015.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Do not answer until the audio ends. Then list each instance of an animal name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.68, 5.54, "canary bird"]], "response_mode": "offline", "answer": "[{\"word\":\"canary bird\",\"start\":4.68,\"end\":5.54}]"} {"utterance_id": "4077-13754-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0006.flac", "task_type": "lexical_detection", "query": "celestial", "prompt": "Once the audio has finished, identify every instance of the word \"celestial\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.67, 2.23, "celestial"], [7.81, 8.35, "celestial"]], "response_mode": "offline", "answer": "[{\"word\":\"celestial\",\"start\":1.67,\"end\":2.23},{\"word\":\"celestial\",\"start\":7.81,\"end\":8.35}]"} {"utterance_id": "1320-122617-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0037.flac", "task_type": "lexical_detection", "query": "dog", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"dog\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.04, 1.39, "dog"]], "response_mode": "offline", "answer": "[{\"word\":\"dog\",\"start\":1.04,\"end\":1.39}]"} {"utterance_id": "3575-170457-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0041.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.51, 5.79, "foot"]], "response_mode": "offline", "answer": "[{\"word\":\"foot\",\"start\":5.51,\"end\":5.79}]"} {"utterance_id": "672-122797-0064", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0064.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen carefully to the entire clip, then report every instance of an ordinal number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.67, 1.23, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":0.67,\"end\":1.23}]"} {"utterance_id": "1320-122617-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0025.flac", "task_type": "lexical_detection", "query": "skin", "prompt": "When the utterance is complete, report all instances of the word \"skin\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.32, 3.63, "skin"]], "response_mode": "offline", "answer": "[{\"word\":\"skin\",\"start\":3.32,\"end\":3.63}]"} {"utterance_id": "7729-102255-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0015.flac", "task_type": "lexical_detection", "query": "outlaws", "prompt": "When the utterance is complete, report all instances of the word \"outlaws\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.31, 5.95, "outlaws"]], "response_mode": "offline", "answer": "[{\"word\":\"outlaws\",\"start\":5.31,\"end\":5.95}]"} {"utterance_id": "5683-32879-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0016.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.24, 1.85, "oracle"]], "response_mode": "offline", "answer": "[{\"word\":\"oracle\",\"start\":1.24,\"end\":1.85}]"} {"utterance_id": "8455-210777-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0012.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.72, 3.04, "son"]], "response_mode": "offline", "answer": "[{\"word\":\"son\",\"start\":2.72,\"end\":3.04}]"} {"utterance_id": "5683-32879-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0018.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.69, 4.27, "dorcas"]], "response_mode": "offline", "answer": "[{\"word\":\"dorcas\",\"start\":3.69,\"end\":4.27}]"} {"utterance_id": "260-123440-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0010.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "When the utterance is complete, report how a body part accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.79, 4.34, "one"], [7.47, 8.05, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.79,\"end\":4.34},{\"word\":\"two\",\"start\":7.47,\"end\":8.05}]"} {"utterance_id": "5105-28240-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0009.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.28, 2.7, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":2.28,\"end\":2.7}]"} {"utterance_id": "5639-40744-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0030.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.96, 1.56, "leocadia"], [13.09, 13.65, "leocadia"]], "response_mode": "offline", "answer": "[{\"word\":\"leocadia\",\"start\":0.96,\"end\":1.56},{\"word\":\"leocadia\",\"start\":13.09,\"end\":13.65}]"} {"utterance_id": "6829-68769-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0044.flac", "task_type": "lexical_detection", "query": "annoyance", "prompt": "Once the audio has finished, identify every instance of the word \"annoyance\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.13, 2.75, "annoyance"]], "response_mode": "offline", "answer": "[{\"word\":\"annoyance\",\"start\":2.13,\"end\":2.75}]"} {"utterance_id": "5142-33396-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0037.flac", "task_type": "lexical_detection", "query": "constant", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"constant\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.58, 2.07, "constant"]], "response_mode": "offline", "answer": "[{\"word\":\"constant\",\"start\":1.58,\"end\":2.07}]"} {"utterance_id": "2830-3980-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0037.flac", "task_type": "lexical_detection", "query": "title", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"title\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.04, 1.42, "title"]], "response_mode": "offline", "answer": "[{\"word\":\"title\",\"start\":1.04,\"end\":1.42}]"} {"utterance_id": "2300-131720-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0008.flac", "task_type": "lexical_detection", "query": "contraction", "prompt": "After listening to the whole recording, output every instance of the word \"contraction\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.91, 5.57, "contraction"]], "response_mode": "offline", "answer": "[{\"word\":\"contraction\",\"start\":4.91,\"end\":5.57}]"} {"utterance_id": "6829-68769-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0052.flac", "task_type": "lexical_detection", "query": "forged", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"forged\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.39, 1.84, "forged"]], "response_mode": "offline", "answer": "[{\"word\":\"forged\",\"start\":1.39,\"end\":1.84}]"} {"utterance_id": "5105-28233-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28233/5105-28233-0006.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Process the full audio offline. Find every instance of a city name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.76, 2.33, "burgos"]], "response_mode": "offline", "answer": "[{\"word\":\"burgos\",\"start\":1.76,\"end\":2.33}]"} {"utterance_id": "4992-41797-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0020.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "After the audio ends, tally every instance of a person's name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.37, 0.85, "one"], [3.8, 4.24, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.37,\"end\":0.85},{\"word\":\"two\",\"start\":3.8,\"end\":4.24}]"} {"utterance_id": "237-134493-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0018.flac", "task_type": "lexical_detection", "query": "row", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"row\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.57, 1.76, "row"]], "response_mode": "offline", "answer": "[{\"word\":\"row\",\"start\":1.57,\"end\":1.76}]"} {"utterance_id": "4992-23283-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0009.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Once the audio has finished, identify every instance of an ordinal number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.87, 5.24, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":4.87,\"end\":5.24}]"} {"utterance_id": "2830-3980-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0026.flac", "task_type": "lexical_detection", "query": "verse", "prompt": "After the audio ends, extract all occurrences of the word \"verse\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 0.87, "verse"]], "response_mode": "offline", "answer": "[{\"word\":\"verse\",\"start\":0.52,\"end\":0.87}]"} {"utterance_id": "8555-284449-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0008.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.86, 4.16, "witch"]], "response_mode": "offline", "answer": "[{\"word\":\"witch\",\"start\":3.86,\"end\":4.16}]"} {"utterance_id": "5683-32865-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0000.flac", "task_type": "lexical_detection", "query": "lake", "prompt": "Once the audio has finished, identify every instance of the word \"lake\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.55, 2.03, "lake"]], "response_mode": "offline", "answer": "[{\"word\":\"lake\",\"start\":1.55,\"end\":2.03}]"} {"utterance_id": "908-157963-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0003.flac", "task_type": "lexical_detection", "query": "fade", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"fade\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.67, 1.12, "fade"]], "response_mode": "offline", "answer": "[{\"word\":\"fade\",\"start\":0.67,\"end\":1.12}]"} {"utterance_id": "121-121726-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0004.flac", "task_type": "lexical_detection", "query": "raised", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"raised\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.77, 3.15, "raised"]], "response_mode": "offline", "answer": "[{\"word\":\"raised\",\"start\":2.77,\"end\":3.15}]"} {"utterance_id": "5105-28240-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0009.flac", "task_type": "lexical_detection", "query": "changing", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"changing\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.86, 1.32, "changing"]], "response_mode": "offline", "answer": "[{\"word\":\"changing\",\"start\":0.86,\"end\":1.32}]"} {"utterance_id": "237-134493-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0000.flac", "task_type": "lexical_detection", "query": "bergson", "prompt": "When the utterance is complete, report all instances of the word \"bergson\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.85, 3.33, "bergson"]], "response_mode": "offline", "answer": "[{\"word\":\"bergson\",\"start\":2.85,\"end\":3.33}]"} {"utterance_id": "4446-2271-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0019.flac", "task_type": "lexical_detection", "query": "actually", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"actually\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.27, 2.83, "actually"]], "response_mode": "offline", "answer": "[{\"word\":\"actually\",\"start\":2.27,\"end\":2.83}]"} {"utterance_id": "1221-135766-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0005.flac", "task_type": "lexical_detection", "query": "child's", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"child's\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.09, 2.53, "child's"]], "response_mode": "offline", "answer": "[{\"word\":\"child's\",\"start\":2.09,\"end\":2.53}]"} {"utterance_id": "4970-29095-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0025.flac", "task_type": "lexical_detection", "query": "won't", "prompt": "Process the full audio offline. Find every instance of the word \"won't\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.16, 1.43, "won't"]], "response_mode": "offline", "answer": "[{\"word\":\"won't\",\"start\":1.16,\"end\":1.43}]"} {"utterance_id": "6829-68769-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0029.flac", "task_type": "lexical_detection", "query": "company", "prompt": "Process the full audio offline. Find every instance of the word \"company\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.87, 1.35, "company"]], "response_mode": "offline", "answer": "[{\"word\":\"company\",\"start\":0.87,\"end\":1.35}]"} {"utterance_id": "4970-29095-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0003.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.25, 2.68, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":2.25,\"end\":2.68}]"} {"utterance_id": "8224-274384-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0005.flac", "task_type": "lexical_detection", "query": "sung", "prompt": "After the audio ends, extract all occurrences of the word \"sung\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.83, 8.36, "sung"]], "response_mode": "offline", "answer": "[{\"word\":\"sung\",\"start\":7.83,\"end\":8.36}]"} {"utterance_id": "5683-32866-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.19, 0.65, "bracton"]], "response_mode": "offline", "answer": "[{\"word\":\"bracton\",\"start\":0.19,\"end\":0.65}]"} {"utterance_id": "2300-131720-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0002.flac", "task_type": "semantic_detection", "query": "city", "prompt": "You will hear a complete utterance. When it ends, list every instance of a city name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.72, 10.25, "new york"]], "response_mode": "offline", "answer": "[{\"word\":\"new york\",\"start\":9.72,\"end\":10.25}]"} {"utterance_id": "1188-133604-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0018.flac", "task_type": "lexical_detection", "query": "failure", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"failure\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.84, 3.22, "failure"]], "response_mode": "offline", "answer": "[{\"word\":\"failure\",\"start\":2.84,\"end\":3.22}]"} {"utterance_id": "2094-142345-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0002.flac", "task_type": "lexical_detection", "query": "solid", "prompt": "When the utterance is complete, report all instances of the word \"solid\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.08, 1.74, "solid"]], "response_mode": "offline", "answer": "[{\"word\":\"solid\",\"start\":1.08,\"end\":1.74}]"} {"utterance_id": "7021-85628-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0017.flac", "task_type": "lexical_detection", "query": "anders", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"anders\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.54, 2.01, "anders"]], "response_mode": "offline", "answer": "[{\"word\":\"anders\",\"start\":1.54,\"end\":2.01}]"} {"utterance_id": "4970-29095-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0020.flac", "task_type": "lexical_detection", "query": "ought", "prompt": "Process the full audio offline. Find every instance of the word \"ought\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.28, 6.5, "ought"]], "response_mode": "offline", "answer": "[{\"word\":\"ought\",\"start\":6.28,\"end\":6.5}]"} {"utterance_id": "3729-6852-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0011.flac", "task_type": "lexical_detection", "query": "name", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"name\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.81, 1.2, "name"]], "response_mode": "offline", "answer": "[{\"word\":\"name\",\"start\":0.81,\"end\":1.2}]"} {"utterance_id": "121-123852-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123852/121-123852-0004.flac", "task_type": "lexical_counting", "query": "doth", "prompt": "Listen to the full audio. Count every occurrence of the word \"doth\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.02, 1.41, "one"], [10.04, 10.45, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.02,\"end\":1.41},{\"word\":\"two\",\"start\":10.04,\"end\":10.45}]"} {"utterance_id": "3570-5696-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0010.flac", "task_type": "lexical_detection", "query": "varying", "prompt": "After the audio ends, extract all occurrences of the word \"varying\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.21, 6.55, "varying"]], "response_mode": "offline", "answer": "[{\"word\":\"varying\",\"start\":6.21,\"end\":6.55}]"} {"utterance_id": "1580-141083-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0034.flac", "task_type": "lexical_detection", "query": "rigidity", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"rigidity\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.83, 3.41, "rigidity"]], "response_mode": "offline", "answer": "[{\"word\":\"rigidity\",\"start\":2.83,\"end\":3.41}]"} {"utterance_id": "5683-32866-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0028.flac", "task_type": "lexical_detection", "query": "hearse", "prompt": "Once the audio has finished, identify every instance of the word \"hearse\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.81, 3.13, "hearse"]], "response_mode": "offline", "answer": "[{\"word\":\"hearse\",\"start\":2.81,\"end\":3.13}]"} {"utterance_id": "237-134500-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0031.flac", "task_type": "lexical_counting", "query": "sometimes", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"sometimes\" (word, start, end). If none, reply [].", "labels": [[0.23, 0.84, "one"], [2.79, 3.31, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.23,\"end\":0.84},{\"word\":\"two\",\"start\":2.79,\"end\":3.31}]"} {"utterance_id": "2300-131720-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0005.flac", "task_type": "lexical_detection", "query": "economy", "prompt": "After the audio ends, extract all occurrences of the word \"economy\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.44, 5.16, "economy"]], "response_mode": "offline", "answer": "[{\"word\":\"economy\",\"start\":4.44,\"end\":5.16}]"} {"utterance_id": "4446-2275-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0000.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a city name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.74, 1.46, "queenstown"]], "response_mode": "offline", "answer": "[{\"word\":\"queenstown\",\"start\":0.74,\"end\":1.46}]"} {"utterance_id": "1320-122612-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0016.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.41, 2.87, "singer's"]], "response_mode": "offline", "answer": "[{\"word\":\"singer's\",\"start\":2.41,\"end\":2.87}]"} {"utterance_id": "237-134493-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0005.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a body part: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[6.19, 6.54, "one"], [7.78, 8.28, "two"], [10.13, 10.62, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.19,\"end\":6.54},{\"word\":\"two\",\"start\":7.78,\"end\":8.28},{\"word\":\"three\",\"start\":10.13,\"end\":10.62}]"} {"utterance_id": "3575-170457-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0014.flac", "task_type": "lexical_detection", "query": "eager", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"eager\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.15, 5.51, "eager"]], "response_mode": "offline", "answer": "[{\"word\":\"eager\",\"start\":5.15,\"end\":5.51}]"} {"utterance_id": "3570-5696-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0009.flac", "task_type": "lexical_detection", "query": "accuracy", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"accuracy\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.74, 1.38, "accuracy"]], "response_mode": "offline", "answer": "[{\"word\":\"accuracy\",\"start\":0.74,\"end\":1.38}]"} {"utterance_id": "237-126133-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0025.flac", "task_type": "lexical_detection", "query": "face", "prompt": "When the utterance is complete, report all instances of the word \"face\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.65, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":2.35,\"end\":2.65}]"} {"utterance_id": "7021-85628-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0010.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Once the audio has finished, identify every instance of a color name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.9, 1.26, "silver"], [4.96, 5.29, "red"]], "response_mode": "offline", "answer": "[{\"word\":\"silver\",\"start\":0.9,\"end\":1.26},{\"word\":\"red\",\"start\":4.96,\"end\":5.29}]"} {"utterance_id": "61-70968-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0041.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Process the full audio offline. For every occurrence of a person's name, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.03, 1.38, "one"], [2.61, 2.81, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.03,\"end\":1.38},{\"word\":\"two\",\"start\":2.61,\"end\":2.81}]"} {"utterance_id": "3570-5695-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0007.flac", "task_type": "semantic_detection", "query": "country", "prompt": "When the utterance is complete, report all instances of a country name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.41, 1.81, "country"]], "response_mode": "offline", "answer": "[{\"word\":\"country\",\"start\":1.41,\"end\":1.81}]"} {"utterance_id": "6829-68769-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0036.flac", "task_type": "lexical_detection", "query": "escort", "prompt": "After listening to the whole recording, output every instance of the word \"escort\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.68, 4.08, "escort"]], "response_mode": "offline", "answer": "[{\"word\":\"escort\",\"start\":3.68,\"end\":4.08}]"} {"utterance_id": "4077-13751-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0021.flac", "task_type": "lexical_detection", "query": "prominent", "prompt": "Process the full audio offline. Find every instance of the word \"prominent\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.1, 10.65, "prominent"]], "response_mode": "offline", "answer": "[{\"word\":\"prominent\",\"start\":10.1,\"end\":10.65}]"} {"utterance_id": "237-126133-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0001.flac", "task_type": "lexical_detection", "query": "fully", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"fully\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[6.82, 7.15, "fully"]], "response_mode": "offline", "answer": "[{\"word\":\"fully\",\"start\":6.82,\"end\":7.15}]"} {"utterance_id": "260-123288-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0014.flac", "task_type": "lexical_detection", "query": "hans", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"hans\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 0.87, "hans"]], "response_mode": "offline", "answer": "[{\"word\":\"hans\",\"start\":0.45,\"end\":0.87}]"} {"utterance_id": "7729-102255-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0005.flac", "task_type": "lexical_detection", "query": "slavery", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"slavery\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.84, 3.4, "slavery"]], "response_mode": "offline", "answer": "[{\"word\":\"slavery\",\"start\":2.84,\"end\":3.4}]"} {"utterance_id": "4507-16021-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0023.flac", "task_type": "semantic_counting", "query": "food", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a food name (word, start, end). If none, reply [].", "labels": [[0.57, 0.91, "one"], [3.0, 3.57, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.57,\"end\":0.91},{\"word\":\"two\",\"start\":3.0,\"end\":3.57}]"} {"utterance_id": "7127-75947-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0036.flac", "task_type": "lexical_detection", "query": "give", "prompt": "Process the full audio offline. Find every instance of the word \"give\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.7, 1.03, "give"]], "response_mode": "offline", "answer": "[{\"word\":\"give\",\"start\":0.7,\"end\":1.03}]"} {"utterance_id": "1089-134686-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0035.flac", "task_type": "lexical_detection", "query": "mountains", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"mountains\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.44, 3.2, "mountains"]], "response_mode": "offline", "answer": "[{\"word\":\"mountains\",\"start\":2.44,\"end\":3.2}]"} {"utterance_id": "4446-2271-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.14, 2.39, "sir harry towne mister bartley alexander"]], "response_mode": "offline", "answer": "[{\"word\":\"sir harry towne mister bartley alexander\",\"start\":0.14,\"end\":2.39}]"} {"utterance_id": "2961-961-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0014.flac", "task_type": "lexical_counting", "query": "founded", "prompt": "After the audio ends, tally every instance of the word \"founded\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[3.64, 4.04, "one"], [6.5, 6.98, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.64,\"end\":4.04},{\"word\":\"two\",\"start\":6.5,\"end\":6.98}]"} {"utterance_id": "1089-134686-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0032.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.3, 2.97, "apostle"]], "response_mode": "offline", "answer": "[{\"word\":\"apostle\",\"start\":2.3,\"end\":2.97}]"} {"utterance_id": "2094-142345-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0005.flac", "task_type": "lexical_detection", "query": "rags", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"rags\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.9, 8.58, "rags"]], "response_mode": "offline", "answer": "[{\"word\":\"rags\",\"start\":7.9,\"end\":8.58}]"} {"utterance_id": "908-31957-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0004.flac", "task_type": "lexical_counting", "query": "home", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"home\" with word/start/end. If none, reply [].", "labels": [[1.31, 1.64, "one"], [13.44, 13.8, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.31,\"end\":1.64},{\"word\":\"two\",\"start\":13.44,\"end\":13.8}]"} {"utterance_id": "7176-92135-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0020.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.72, 1.98, "nine two three"], [3.6, 3.81, "nine"]], "response_mode": "offline", "answer": "[{\"word\":\"nine two three\",\"start\":0.72,\"end\":1.98},{\"word\":\"nine\",\"start\":3.6,\"end\":3.81}]"} {"utterance_id": "61-70970-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0035.flac", "task_type": "lexical_detection", "query": "estate", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"estate\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.59, 5.97, "estate"]], "response_mode": "offline", "answer": "[{\"word\":\"estate\",\"start\":5.59,\"end\":5.97}]"} {"utterance_id": "7021-79730-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0003.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "Once the audio has finished, identify every instance of a disease name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[31.95, 32.66, "injury"]], "response_mode": "offline", "answer": "[{\"word\":\"injury\",\"start\":31.95,\"end\":32.66}]"} {"utterance_id": "2961-960-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0022.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.48, 0.93, "plato"]], "response_mode": "offline", "answer": "[{\"word\":\"plato\",\"start\":0.48,\"end\":0.93}]"} {"utterance_id": "7127-75947-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0027.flac", "task_type": "lexical_detection", "query": "athenais", "prompt": "After listening to the whole recording, output every instance of the word \"athenais\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.01, 5.66, "athenais"]], "response_mode": "offline", "answer": "[{\"word\":\"athenais\",\"start\":5.01,\"end\":5.66}]"} {"utterance_id": "5683-32879-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.92, 1.71, "rachel lake"]], "response_mode": "offline", "answer": "[{\"word\":\"rachel lake\",\"start\":0.92,\"end\":1.71}]"} {"utterance_id": "5142-33396-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0003.flac", "task_type": "lexical_detection", "query": "ships", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"ships\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.7, 3.28, "ships"]], "response_mode": "offline", "answer": "[{\"word\":\"ships\",\"start\":2.7,\"end\":3.28}]"} {"utterance_id": "1089-134686-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0031.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.02, 1.4, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":1.02,\"end\":1.4}]"} {"utterance_id": "1320-122617-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0004.flac", "task_type": "lexical_detection", "query": "ingenious", "prompt": "When the utterance is complete, report all instances of the word \"ingenious\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.67, 1.2, "ingenious"]], "response_mode": "offline", "answer": "[{\"word\":\"ingenious\",\"start\":0.67,\"end\":1.2}]"} {"utterance_id": "6930-81414-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0006.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.88, 2.37, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":1.88,\"end\":2.37}]"} {"utterance_id": "1221-135767-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0015.flac", "task_type": "lexical_detection", "query": "worship", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"worship\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.48, 1.97, "worship"]], "response_mode": "offline", "answer": "[{\"word\":\"worship\",\"start\":1.48,\"end\":1.97}]"} {"utterance_id": "5639-40744-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0033.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.65, 5.82, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":5.65,\"end\":5.82}]"} {"utterance_id": "5142-36377-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0018.flac", "task_type": "lexical_counting", "query": "looked", "prompt": "Process the full audio offline. For every occurrence of the word \"looked\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.36, 0.64, "one"], [2.82, 3.04, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.36,\"end\":0.64},{\"word\":\"two\",\"start\":2.82,\"end\":3.04}]"} {"utterance_id": "908-157963-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0030.flac", "task_type": "lexical_detection", "query": "whirlpool", "prompt": "Process the full audio offline. Find every instance of the word \"whirlpool\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.46, 2.28, "whirlpool"]], "response_mode": "offline", "answer": "[{\"word\":\"whirlpool\",\"start\":1.46,\"end\":2.28}]"} {"utterance_id": "6930-81414-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0008.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.49, 4.91, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":4.49,\"end\":4.91}]"} {"utterance_id": "8455-210777-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0008.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.9, 6.1, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.9,\"end\":6.1}]"} {"utterance_id": "1995-1826-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0011.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.82, 2.39, "children"]], "response_mode": "offline", "answer": "[{\"word\":\"children\",\"start\":1.82,\"end\":2.39}]"} {"utterance_id": "1995-1836-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.19, 2.84, "cresswell"]], "response_mode": "offline", "answer": "[{\"word\":\"cresswell\",\"start\":2.19,\"end\":2.84}]"} {"utterance_id": "8463-294828-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0029.flac", "task_type": "lexical_detection", "query": "frigate", "prompt": "When the utterance is complete, report all instances of the word \"frigate\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.12, 3.63, "frigate"]], "response_mode": "offline", "answer": "[{\"word\":\"frigate\",\"start\":3.12,\"end\":3.63}]"} {"utterance_id": "1188-133604-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0028.flac", "task_type": "lexical_detection", "query": "highest", "prompt": "Once the audio has finished, identify every instance of the word \"highest\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.27, 4.67, "highest"]], "response_mode": "offline", "answer": "[{\"word\":\"highest\",\"start\":4.27,\"end\":4.67}]"} {"utterance_id": "7127-75947-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0038.flac", "task_type": "lexical_detection", "query": "space", "prompt": "When the utterance is complete, report all instances of the word \"space\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.71, 8.21, "space"]], "response_mode": "offline", "answer": "[{\"word\":\"space\",\"start\":7.71,\"end\":8.21}]"} {"utterance_id": "4992-41797-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0009.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.65, 5.28, "professor"]], "response_mode": "offline", "answer": "[{\"word\":\"professor\",\"start\":4.65,\"end\":5.28}]"} {"utterance_id": "5683-32866-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0020.flac", "task_type": "lexical_detection", "query": "showed", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"showed\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.15, 2.5, "showed"]], "response_mode": "offline", "answer": "[{\"word\":\"showed\",\"start\":2.15,\"end\":2.5}]"} {"utterance_id": "6829-68771-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0020.flac", "task_type": "lexical_detection", "query": "parsons", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"parsons\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.7, 4.27, "parsons"]], "response_mode": "offline", "answer": "[{\"word\":\"parsons\",\"start\":3.7,\"end\":4.27}]"} {"utterance_id": "5683-32865-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.95, 1.31, "lake"]], "response_mode": "offline", "answer": "[{\"word\":\"lake\",\"start\":0.95,\"end\":1.31}]"} {"utterance_id": "7021-79759-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79759/7021-79759-0003.flac", "task_type": "lexical_detection", "query": "importance", "prompt": "Once the audio has finished, identify every instance of the word \"importance\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.09, 1.76, "importance"]], "response_mode": "offline", "answer": "[{\"word\":\"importance\",\"start\":1.09,\"end\":1.76}]"} {"utterance_id": "7127-75947-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0001.flac", "task_type": "lexical_detection", "query": "observing", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"observing\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.77, 6.62, "observing"]], "response_mode": "offline", "answer": "[{\"word\":\"observing\",\"start\":5.77,\"end\":6.62}]"} {"utterance_id": "908-157963-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0020.flac", "task_type": "semantic_detection", "query": "food", "prompt": "After listening to the whole recording, output every instance of a food name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.14, 7.65, "food"]], "response_mode": "offline", "answer": "[{\"word\":\"food\",\"start\":7.14,\"end\":7.65}]"} {"utterance_id": "4446-2275-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0023.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.55, 1.14, "alexander"]], "response_mode": "offline", "answer": "[{\"word\":\"alexander\",\"start\":0.55,\"end\":1.14}]"} {"utterance_id": "1089-134691-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0008.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.01, 1.49, "faces"], [14.05, 14.55, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"faces\",\"start\":1.01,\"end\":1.49},{\"word\":\"face\",\"start\":14.05,\"end\":14.55}]"} {"utterance_id": "4077-13754-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0016.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, list every instance of an organization name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.31, 1.82, "mormonism"]], "response_mode": "offline", "answer": "[{\"word\":\"mormonism\",\"start\":1.31,\"end\":1.82}]"} {"utterance_id": "237-134500-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0002.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After listening to the whole recording, output every instance of a color name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.91, 3.26, "white"]], "response_mode": "offline", "answer": "[{\"word\":\"white\",\"start\":2.91,\"end\":3.26}]"} {"utterance_id": "5639-40744-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0006.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.48, 1.13, "rodolfo"], [4.43, 4.95, "leocadia"]], "response_mode": "offline", "answer": "[{\"word\":\"rodolfo\",\"start\":0.48,\"end\":1.13},{\"word\":\"leocadia\",\"start\":4.43,\"end\":4.95}]"} {"utterance_id": "1284-1180-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0008.flac", "task_type": "lexical_detection", "query": "noon", "prompt": "Once the audio has finished, identify every instance of the word \"noon\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.15, 3.57, "noon"]], "response_mode": "offline", "answer": "[{\"word\":\"noon\",\"start\":3.15,\"end\":3.57}]"} {"utterance_id": "7127-75946-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0016.flac", "task_type": "lexical_detection", "query": "seasons", "prompt": "When the utterance is complete, report all instances of the word \"seasons\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.56, 1.29, "seasons"]], "response_mode": "offline", "answer": "[{\"word\":\"seasons\",\"start\":0.56,\"end\":1.29}]"} {"utterance_id": "5639-40744-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0033.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Once the audio has finished, identify every instance of a family relation. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.24, 3.58, "son"]], "response_mode": "offline", "answer": "[{\"word\":\"son\",\"start\":3.24,\"end\":3.58}]"} {"utterance_id": "1089-134691-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.4, 2.69, "brother mac ardle brother keogh"]], "response_mode": "offline", "answer": "[{\"word\":\"brother mac ardle brother keogh\",\"start\":0.4,\"end\":2.69}]"} {"utterance_id": "6930-76324-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.37, 1.0, "goliath"]], "response_mode": "offline", "answer": "[{\"word\":\"goliath\",\"start\":0.37,\"end\":1.0}]"} {"utterance_id": "5639-40744-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0041.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a person's name: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.87, 1.32, "one"], [7.14, 7.77, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.87,\"end\":1.32},{\"word\":\"two\",\"start\":7.14,\"end\":7.77}]"} {"utterance_id": "5105-28241-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0012.flac", "task_type": "lexical_detection", "query": "aloud", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"aloud\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.51, 3.04, "aloud"]], "response_mode": "offline", "answer": "[{\"word\":\"aloud\",\"start\":2.51,\"end\":3.04}]"} {"utterance_id": "5142-36586-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36586/5142-36586-0001.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.41, 2.01, "animals"]], "response_mode": "offline", "answer": "[{\"word\":\"animals\",\"start\":1.41,\"end\":2.01}]"} {"utterance_id": "5683-32879-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.36, 1.18, "mark wylder"]], "response_mode": "offline", "answer": "[{\"word\":\"mark wylder\",\"start\":0.36,\"end\":1.18}]"} {"utterance_id": "260-123286-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0024.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.28, 2.89, "professor"]], "response_mode": "offline", "answer": "[{\"word\":\"professor\",\"start\":2.28,\"end\":2.89}]"} {"utterance_id": "61-70970-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0017.flac", "task_type": "lexical_detection", "query": "rest", "prompt": "After the audio ends, extract all occurrences of the word \"rest\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.26, 0.67, "rest"]], "response_mode": "offline", "answer": "[{\"word\":\"rest\",\"start\":0.26,\"end\":0.67}]"} {"utterance_id": "7729-102255-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0044.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a city name with fields word, start, end (seconds). If none, reply [].", "labels": [[14.75, 15.43, "leavenworth"]], "response_mode": "offline", "answer": "[{\"word\":\"leavenworth\",\"start\":14.75,\"end\":15.43}]"} {"utterance_id": "2094-142345-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0000.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[18.65, 18.9, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":18.65,\"end\":18.9}]"} {"utterance_id": "260-123288-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0016.flac", "task_type": "lexical_detection", "query": "refer", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"refer\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.27, 0.68, "refer"]], "response_mode": "offline", "answer": "[{\"word\":\"refer\",\"start\":0.27,\"end\":0.68}]"} {"utterance_id": "1188-133604-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0028.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.96, 1.57, "last"], [8.86, 9.18, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":0.96,\"end\":1.57},{\"word\":\"first\",\"start\":8.86,\"end\":9.18}]"} {"utterance_id": "8463-294828-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0028.flac", "task_type": "lexical_detection", "query": "credit", "prompt": "Once the audio has finished, identify every instance of the word \"credit\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.15, 1.55, "credit"]], "response_mode": "offline", "answer": "[{\"word\":\"credit\",\"start\":1.15,\"end\":1.55}]"} {"utterance_id": "6930-76324-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0022.flac", "task_type": "lexical_detection", "query": "remarked", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"remarked\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.37, 1.8, "remarked"]], "response_mode": "offline", "answer": "[{\"word\":\"remarked\",\"start\":1.37,\"end\":1.8}]"} {"utterance_id": "4077-13751-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0009.flac", "task_type": "lexical_detection", "query": "temple", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"temple\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.44, 10.76, "temple"]], "response_mode": "offline", "answer": "[{\"word\":\"temple\",\"start\":10.44,\"end\":10.76}]"} {"utterance_id": "4992-41806-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0008.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After listening to the whole recording, output every instance of a color name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.41, 2.66, "yellow"]], "response_mode": "offline", "answer": "[{\"word\":\"yellow\",\"start\":2.41,\"end\":2.66}]"} {"utterance_id": "2961-961-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0015.flac", "task_type": "lexical_detection", "query": "exist", "prompt": "After listening to the whole recording, output every instance of the word \"exist\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.14, 1.59, "exist"]], "response_mode": "offline", "answer": "[{\"word\":\"exist\",\"start\":1.14,\"end\":1.59}]"} {"utterance_id": "4077-13751-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0021.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.57, 6.15, "officers"]], "response_mode": "offline", "answer": "[{\"word\":\"officers\",\"start\":5.57,\"end\":6.15}]"} {"utterance_id": "3729-6852-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0019.flac", "task_type": "lexical_detection", "query": "coffee", "prompt": "Once the audio has finished, identify every instance of the word \"coffee\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.57, 2.09, "coffee"]], "response_mode": "offline", "answer": "[{\"word\":\"coffee\",\"start\":1.57,\"end\":2.09}]"} {"utterance_id": "4970-29095-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0002.flac", "task_type": "lexical_detection", "query": "shade", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"shade\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.86, 4.12, "shade"]], "response_mode": "offline", "answer": "[{\"word\":\"shade\",\"start\":3.86,\"end\":4.12}]"} {"utterance_id": "6930-76324-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0012.flac", "task_type": "lexical_detection", "query": "cleaning", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"cleaning\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.13, 4.61, "cleaning"]], "response_mode": "offline", "answer": "[{\"word\":\"cleaning\",\"start\":4.13,\"end\":4.61}]"} {"utterance_id": "8455-210777-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0020.flac", "task_type": "lexical_detection", "query": "jack", "prompt": "Once the audio has finished, identify every instance of the word \"jack\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.26, 1.77, "jack"]], "response_mode": "offline", "answer": "[{\"word\":\"jack\",\"start\":1.26,\"end\":1.77}]"} {"utterance_id": "1221-135767-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0007.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen carefully to the entire clip, then report every instance of a color name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.12, 10.63, "scarlet"]], "response_mode": "offline", "answer": "[{\"word\":\"scarlet\",\"start\":10.12,\"end\":10.63}]"} {"utterance_id": "7021-85628-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0006.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.89, 3.38, "anders"]], "response_mode": "offline", "answer": "[{\"word\":\"anders\",\"start\":2.89,\"end\":3.38}]"} {"utterance_id": "260-123288-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0013.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Process the full audio offline. Find every instance of a food name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.01, 2.5, "water"]], "response_mode": "offline", "answer": "[{\"word\":\"water\",\"start\":2.01,\"end\":2.5}]"} {"utterance_id": "2300-131720-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0026.flac", "task_type": "lexical_detection", "query": "arc", "prompt": "After listening to the whole recording, output every instance of the word \"arc\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.46, 0.72, "arc"]], "response_mode": "offline", "answer": "[{\"word\":\"arc\",\"start\":0.46,\"end\":0.72}]"} {"utterance_id": "1284-1180-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0021.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.86, 2.33, "magician"]], "response_mode": "offline", "answer": "[{\"word\":\"magician\",\"start\":1.86,\"end\":2.33}]"} {"utterance_id": "2830-3980-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0032.flac", "task_type": "lexical_detection", "query": "hath", "prompt": "Process the full audio offline. Find every instance of the word \"hath\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.03, 2.29, "hath"], [5.04, 5.28, "hath"]], "response_mode": "offline", "answer": "[{\"word\":\"hath\",\"start\":2.03,\"end\":2.29},{\"word\":\"hath\",\"start\":5.04,\"end\":5.28}]"} {"utterance_id": "4970-29093-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0022.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the full audio. After it finishes, find every instance of an ordinal number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.15, 3.6, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":3.15,\"end\":3.6}]"} {"utterance_id": "1320-122612-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0012.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Process the full audio offline. Find every instance of a plant name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.52, 6.9, "trees"]], "response_mode": "offline", "answer": "[{\"word\":\"trees\",\"start\":6.52,\"end\":6.9}]"} {"utterance_id": "2300-131720-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0016.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.0, 6.71, "mister edison"]], "response_mode": "offline", "answer": "[{\"word\":\"mister edison\",\"start\":6.0,\"end\":6.71}]"} {"utterance_id": "4077-13751-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0004.flac", "task_type": "semantic_counting", "query": "organization", "prompt": "Once the audio has finished, list each occurrence of an organization name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[9.75, 10.57, "one"], [15.83, 16.27, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":9.75,\"end\":10.57},{\"word\":\"two\",\"start\":15.83,\"end\":16.27}]"} {"utterance_id": "6930-76324-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0021.flac", "task_type": "lexical_detection", "query": "hint", "prompt": "Once the audio has finished, identify every instance of the word \"hint\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.6, 5.07, "hint"]], "response_mode": "offline", "answer": "[{\"word\":\"hint\",\"start\":4.6,\"end\":5.07}]"} {"utterance_id": "8463-294825-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0017.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.31, 1.59, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.31,\"end\":1.59}]"} {"utterance_id": "1188-133604-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0004.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen carefully to the entire clip, then report every instance of a color name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.57, 2.85, "tint"]], "response_mode": "offline", "answer": "[{\"word\":\"tint\",\"start\":2.57,\"end\":2.85}]"} {"utterance_id": "7729-102255-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0040.flac", "task_type": "lexical_detection", "query": "factions", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"factions\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.66, 5.38, "factions"]], "response_mode": "offline", "answer": "[{\"word\":\"factions\",\"start\":4.66,\"end\":5.38}]"} {"utterance_id": "8463-294828-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0015.flac", "task_type": "lexical_detection", "query": "appeared", "prompt": "After the audio ends, extract all occurrences of the word \"appeared\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.99, 4.61, "appeared"]], "response_mode": "offline", "answer": "[{\"word\":\"appeared\",\"start\":3.99,\"end\":4.61}]"} {"utterance_id": "7021-79730-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0002.flac", "task_type": "lexical_detection", "query": "affection", "prompt": "After the audio ends, extract all occurrences of the word \"affection\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.7, 2.36, "affection"]], "response_mode": "offline", "answer": "[{\"word\":\"affection\",\"start\":1.7,\"end\":2.36}]"} {"utterance_id": "2300-131720-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0029.flac", "task_type": "lexical_detection", "query": "electrolytic", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"electrolytic\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.09, 1.86, "electrolytic"]], "response_mode": "offline", "answer": "[{\"word\":\"electrolytic\",\"start\":1.09,\"end\":1.86}]"} {"utterance_id": "2961-960-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0000.flac", "task_type": "lexical_counting", "query": "numbers", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"numbers\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[4.32, 4.97, "one"], [6.52, 7.03, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.32,\"end\":4.97},{\"word\":\"two\",\"start\":6.52,\"end\":7.03}]"} {"utterance_id": "260-123288-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0000.flac", "task_type": "lexical_detection", "query": "distance", "prompt": "After listening to the whole recording, output every instance of the word \"distance\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.89, 2.57, "distance"]], "response_mode": "offline", "answer": "[{\"word\":\"distance\",\"start\":1.89,\"end\":2.57}]"} {"utterance_id": "4970-29095-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0037.flac", "task_type": "lexical_counting", "query": "understood", "prompt": "After listening to the whole recording, output timed count words for the word \"understood\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[2.3, 2.86, "one"], [5.24, 5.78, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.3,\"end\":2.86},{\"word\":\"two\",\"start\":5.24,\"end\":5.78}]"} {"utterance_id": "2830-3980-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0032.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.81, 2.03, "eye"], [3.66, 3.86, "ear"], [5.98, 6.27, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"eye\",\"start\":1.81,\"end\":2.03},{\"word\":\"ear\",\"start\":3.66,\"end\":3.86},{\"word\":\"heart\",\"start\":5.98,\"end\":6.27}]"} {"utterance_id": "1089-134691-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0015.flac", "task_type": "lexical_detection", "query": "words", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"words\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.39, 1.07, "words"]], "response_mode": "offline", "answer": "[{\"word\":\"words\",\"start\":0.39,\"end\":1.07}]"} {"utterance_id": "1089-134686-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0028.flac", "task_type": "semantic_counting", "query": "date", "prompt": "When the utterance is complete, report how a date accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.89, 2.47, "one"], [6.72, 7.43, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.89,\"end\":2.47},{\"word\":\"two\",\"start\":6.72,\"end\":7.43}]"} {"utterance_id": "3729-6852-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0033.flac", "task_type": "lexical_counting", "query": "silliness", "prompt": "Once the audio has finished, list each occurrence of the word \"silliness\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[19.64, 20.3, "one"], [25.5, 26.29, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":19.64,\"end\":20.3},{\"word\":\"two\",\"start\":25.5,\"end\":26.29}]"} {"utterance_id": "6829-68769-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.44, 3.77, "kenneth"]], "response_mode": "offline", "answer": "[{\"word\":\"kenneth\",\"start\":3.44,\"end\":3.77}]"} {"utterance_id": "5683-32866-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0007.flac", "task_type": "lexical_detection", "query": "he's", "prompt": "After the audio ends, extract all occurrences of the word \"he's\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.22, 2.62, "he's"]], "response_mode": "offline", "answer": "[{\"word\":\"he's\",\"start\":2.22,\"end\":2.62}]"} {"utterance_id": "8555-292519-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0012.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen carefully to the entire clip, then report every instance of a color name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.96, 1.49, "black"]], "response_mode": "offline", "answer": "[{\"word\":\"black\",\"start\":0.96,\"end\":1.49}]"} {"utterance_id": "1995-1837-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0025.flac", "task_type": "lexical_detection", "query": "gathered", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"gathered\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.94, 2.43, "gathered"]], "response_mode": "offline", "answer": "[{\"word\":\"gathered\",\"start\":1.94,\"end\":2.43}]"} {"utterance_id": "3575-170457-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0019.flac", "task_type": "lexical_detection", "query": "reply", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"reply\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.99, 2.51, "reply"]], "response_mode": "offline", "answer": "[{\"word\":\"reply\",\"start\":1.99,\"end\":2.51}]"} {"utterance_id": "5142-33396-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0001.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.02, 1.42, "olaf"]], "response_mode": "offline", "answer": "[{\"word\":\"olaf\",\"start\":1.02,\"end\":1.42}]"} {"utterance_id": "8463-294828-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0036.flac", "task_type": "lexical_detection", "query": "curiosity", "prompt": "Once the audio has finished, identify every instance of the word \"curiosity\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.38, 6.12, "curiosity"]], "response_mode": "offline", "answer": "[{\"word\":\"curiosity\",\"start\":5.38,\"end\":6.12}]"} {"utterance_id": "2094-142345-0057", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0057.flac", "task_type": "lexical_detection", "query": "smiling", "prompt": "When the utterance is complete, report all instances of the word \"smiling\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.51, 7.02, "smiling"]], "response_mode": "offline", "answer": "[{\"word\":\"smiling\",\"start\":6.51,\"end\":7.02}]"} {"utterance_id": "1580-141083-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0035.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.31, 0.64, "holmes"]], "response_mode": "offline", "answer": "[{\"word\":\"holmes\",\"start\":0.31,\"end\":0.64}]"} {"utterance_id": "4970-29095-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0008.flac", "task_type": "lexical_detection", "query": "mother", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"mother\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.34, 0.82, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":0.34,\"end\":0.82}]"} {"utterance_id": "3570-5696-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0001.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Once the audio has finished, identify every instance of an ordinal number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.46, 5.78, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":5.46,\"end\":5.78}]"} {"utterance_id": "6829-68769-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0047.flac", "task_type": "lexical_detection", "query": "marshall", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"marshall\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.72, 2.22, "marshall"]], "response_mode": "offline", "answer": "[{\"word\":\"marshall\",\"start\":1.72,\"end\":2.22}]"} {"utterance_id": "121-127105-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.27, 2.86, "douglas"]], "response_mode": "offline", "answer": "[{\"word\":\"douglas\",\"start\":2.27,\"end\":2.86}]"} {"utterance_id": "1221-135766-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0007.flac", "task_type": "lexical_detection", "query": "risk", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"risk\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.7, 6.06, "risk"]], "response_mode": "offline", "answer": "[{\"word\":\"risk\",\"start\":5.7,\"end\":6.06}]"} {"utterance_id": "8555-284449-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0017.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.81, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":0.53,\"end\":0.81}]"} {"utterance_id": "7127-75946-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0020.flac", "task_type": "lexical_detection", "query": "retained", "prompt": "Process the full audio offline. Find every instance of the word \"retained\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.62, 6.03, "retained"]], "response_mode": "offline", "answer": "[{\"word\":\"retained\",\"start\":5.62,\"end\":6.03}]"} {"utterance_id": "672-122797-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0040.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "After the audio ends, extract all occurrences of a time of day. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.89, 1.33, "night"]], "response_mode": "offline", "answer": "[{\"word\":\"night\",\"start\":0.89,\"end\":1.33}]"} {"utterance_id": "1284-134647-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0003.flac", "task_type": "lexical_detection", "query": "heretics", "prompt": "After the audio ends, extract all occurrences of the word \"heretics\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.33, 3.11, "heretics"]], "response_mode": "offline", "answer": "[{\"word\":\"heretics\",\"start\":2.33,\"end\":3.11}]"} {"utterance_id": "237-126133-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0009.flac", "task_type": "lexical_detection", "query": "polly", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"polly\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.48, 2.92, "polly"]], "response_mode": "offline", "answer": "[{\"word\":\"polly\",\"start\":2.48,\"end\":2.92}]"} {"utterance_id": "6829-68769-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0023.flac", "task_type": "lexical_detection", "query": "foolish", "prompt": "Process the full audio offline. Find every instance of the word \"foolish\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.99, 2.44, "foolish"]], "response_mode": "offline", "answer": "[{\"word\":\"foolish\",\"start\":1.99,\"end\":2.44}]"} {"utterance_id": "6829-68769-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0007.flac", "task_type": "lexical_detection", "query": "circumstances", "prompt": "After listening to the whole recording, output every instance of the word \"circumstances\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.67, 1.58, "circumstances"]], "response_mode": "offline", "answer": "[{\"word\":\"circumstances\",\"start\":0.67,\"end\":1.58}]"} {"utterance_id": "8224-274384-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0010.flac", "task_type": "lexical_detection", "query": "settlement", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"settlement\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.97, 1.55, "settlement"]], "response_mode": "offline", "answer": "[{\"word\":\"settlement\",\"start\":0.97,\"end\":1.55}]"} {"utterance_id": "8463-294825-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0009.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Do not answer until the audio ends. Then list each instance of a state or province name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[18.9, 19.74, "south pole"]], "response_mode": "offline", "answer": "[{\"word\":\"south pole\",\"start\":18.9,\"end\":19.74}]"} {"utterance_id": "3575-170457-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0003.flac", "task_type": "lexical_detection", "query": "loving", "prompt": "When the utterance is complete, report all instances of the word \"loving\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.32, 2.68, "loving"]], "response_mode": "offline", "answer": "[{\"word\":\"loving\",\"start\":2.32,\"end\":2.68}]"} {"utterance_id": "4970-29095-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0026.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the full audio. After it finishes, find every instance of a plant name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.99, 8.58, "trees"], [10.12, 10.66, "leaves"]], "response_mode": "offline", "answer": "[{\"word\":\"trees\",\"start\":7.99,\"end\":8.58},{\"word\":\"leaves\",\"start\":10.12,\"end\":10.66}]"} {"utterance_id": "4992-41806-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0000.flac", "task_type": "lexical_detection", "query": "natty", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"natty\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 0.89, "natty"]], "response_mode": "offline", "answer": "[{\"word\":\"natty\",\"start\":0.42,\"end\":0.89}]"} {"utterance_id": "1221-135766-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0011.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.85, 4.42, "child"]], "response_mode": "offline", "answer": "[{\"word\":\"child\",\"start\":3.85,\"end\":4.42}]"} {"utterance_id": "1089-134691-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0011.flac", "task_type": "lexical_detection", "query": "names", "prompt": "After listening to the whole recording, output every instance of the word \"names\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.63, 3.1, "names"]], "response_mode": "offline", "answer": "[{\"word\":\"names\",\"start\":2.63,\"end\":3.1}]"} {"utterance_id": "1284-1180-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0021.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.82, 1.12, "next"]], "response_mode": "offline", "answer": "[{\"word\":\"next\",\"start\":0.82,\"end\":1.12}]"} {"utterance_id": "4446-2275-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0029.flac", "task_type": "lexical_counting", "query": "tell", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"tell\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.63, 0.79, "one"], [2.9, 3.15, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.63,\"end\":0.79},{\"word\":\"two\",\"start\":2.9,\"end\":3.15}]"} {"utterance_id": "1995-1826-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0001.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Once the audio has finished, identify every instance of a state or province name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.37, 0.85, "south"]], "response_mode": "offline", "answer": "[{\"word\":\"south\",\"start\":0.37,\"end\":0.85}]"} {"utterance_id": "61-70970-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0010.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.7, 1.15, "squire"]], "response_mode": "offline", "answer": "[{\"word\":\"squire\",\"start\":0.7,\"end\":1.15}]"} {"utterance_id": "7176-88083-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0027.flac", "task_type": "lexical_detection", "query": "parted", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"parted\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.85, 1.35, "parted"]], "response_mode": "offline", "answer": "[{\"word\":\"parted\",\"start\":0.85,\"end\":1.35}]"} {"utterance_id": "8555-284447-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0003.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.48, 1.0, "cap'n bill"]], "response_mode": "offline", "answer": "[{\"word\":\"cap'n bill\",\"start\":0.48,\"end\":1.0}]"} {"utterance_id": "1221-135767-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0024.flac", "task_type": "lexical_counting", "query": "rose", "prompt": "After listening to the whole recording, output timed count words for the word \"rose\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[1.37, 1.66, "one"], [3.5, 3.92, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.37,\"end\":1.66},{\"word\":\"two\",\"start\":3.5,\"end\":3.92}]"} {"utterance_id": "1221-135767-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0006.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "You will hear a complete utterance. When it ends, list every instance of a family relation you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.48, 0.77, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":0.48,\"end\":0.77}]"} {"utterance_id": "2961-960-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0009.flac", "task_type": "lexical_detection", "query": "creation", "prompt": "When the utterance is complete, report all instances of the word \"creation\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.66, 4.3, "creation"]], "response_mode": "offline", "answer": "[{\"word\":\"creation\",\"start\":3.66,\"end\":4.3}]"} {"utterance_id": "1284-1181-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0014.flac", "task_type": "lexical_detection", "query": "box", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"box\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.88, 3.24, "box"]], "response_mode": "offline", "answer": "[{\"word\":\"box\",\"start\":2.88,\"end\":3.24}]"} {"utterance_id": "2094-142345-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0030.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 0.9, "totty"]], "response_mode": "offline", "answer": "[{\"word\":\"totty\",\"start\":0.51,\"end\":0.9}]"} {"utterance_id": "3575-170457-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0031.flac", "task_type": "semantic_detection", "query": "date", "prompt": "Listen carefully to the entire clip, then report every instance of a date as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 2.75, "august twenty seventh eighteen thirty seven"]], "response_mode": "offline", "answer": "[{\"word\":\"august twenty seventh eighteen thirty seven\",\"start\":0.51,\"end\":2.75}]"} {"utterance_id": "5683-32879-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0022.flac", "task_type": "lexical_counting", "query": "like", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"like\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.69, 0.95, "one"], [3.33, 3.63, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.69,\"end\":0.95},{\"word\":\"two\",\"start\":3.33,\"end\":3.63}]"} {"utterance_id": "260-123440-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0014.flac", "task_type": "lexical_detection", "query": "bad", "prompt": "Once the audio has finished, identify every instance of the word \"bad\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.02, 2.5, "bad"]], "response_mode": "offline", "answer": "[{\"word\":\"bad\",\"start\":2.02,\"end\":2.5}]"} {"utterance_id": "7176-92135-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0044.flac", "task_type": "lexical_detection", "query": "cigarette", "prompt": "After the audio ends, extract all occurrences of the word \"cigarette\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.76, 1.27, "cigarette"]], "response_mode": "offline", "answer": "[{\"word\":\"cigarette\",\"start\":0.76,\"end\":1.27}]"} {"utterance_id": "6829-68769-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0042.flac", "task_type": "lexical_detection", "query": "sir", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"sir\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.02, 1.28, "sir"]], "response_mode": "offline", "answer": "[{\"word\":\"sir\",\"start\":1.02,\"end\":1.28}]"} {"utterance_id": "4992-41797-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0022.flac", "task_type": "lexical_detection", "query": "older", "prompt": "After listening to the whole recording, output every instance of the word \"older\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 0.76, "older"]], "response_mode": "offline", "answer": "[{\"word\":\"older\",\"start\":0.52,\"end\":0.76}]"} {"utterance_id": "260-123440-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0005.flac", "task_type": "lexical_detection", "query": "yesterday", "prompt": "Process the full audio offline. Find every instance of the word \"yesterday\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.37, 1.01, "yesterday"]], "response_mode": "offline", "answer": "[{\"word\":\"yesterday\",\"start\":0.37,\"end\":1.01}]"} {"utterance_id": "4507-16021-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0023.flac", "task_type": "lexical_detection", "query": "manufacturer", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"manufacturer\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.91, 2.05, "manufacturer"], [9.5, 10.61, "manufacturer"]], "response_mode": "offline", "answer": "[{\"word\":\"manufacturer\",\"start\":0.91,\"end\":2.05},{\"word\":\"manufacturer\",\"start\":9.5,\"end\":10.61}]"} {"utterance_id": "2094-142345-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0013.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Listen to the full audio. Count every occurrence of a body part. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.64, 1.07, "one"], [2.52, 2.99, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.64,\"end\":1.07},{\"word\":\"two\",\"start\":2.52,\"end\":2.99}]"} {"utterance_id": "61-70968-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0027.flac", "task_type": "lexical_detection", "query": "themselves", "prompt": "When the utterance is complete, report all instances of the word \"themselves\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.02, 6.65, "themselves"]], "response_mode": "offline", "answer": "[{\"word\":\"themselves\",\"start\":6.02,\"end\":6.65}]"} {"utterance_id": "8463-294828-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0005.flac", "task_type": "lexical_detection", "query": "conseil", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"conseil\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.23, 0.89, "conseil"]], "response_mode": "offline", "answer": "[{\"word\":\"conseil\",\"start\":0.23,\"end\":0.89}]"} {"utterance_id": "2094-142345-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0029.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After the audio ends, extract all occurrences of a color name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.14, 7.41, "blue"]], "response_mode": "offline", "answer": "[{\"word\":\"blue\",\"start\":7.14,\"end\":7.41}]"} {"utterance_id": "3575-170457-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0005.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[1.33, 1.92, "clergyman's"]], "response_mode": "offline", "answer": "[{\"word\":\"clergyman's\",\"start\":1.33,\"end\":1.92}]"} {"utterance_id": "7176-88083-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0002.flac", "task_type": "semantic_counting", "query": "color", "prompt": "When the utterance is complete, report how a color name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.96, 1.51, "one"], [6.86, 7.24, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.96,\"end\":1.51},{\"word\":\"two\",\"start\":6.86,\"end\":7.24}]"} {"utterance_id": "1995-1826-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0005.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.46, 1.16, "john"]], "response_mode": "offline", "answer": "[{\"word\":\"john\",\"start\":0.46,\"end\":1.16}]"} {"utterance_id": "5142-36377-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0023.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the full audio. Count every occurrence of a person's name. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.97, 2.36, "one"], [3.6, 4.38, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.97,\"end\":2.36},{\"word\":\"two\",\"start\":3.6,\"end\":4.38}]"} {"utterance_id": "7127-75947-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0009.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[22.03, 22.44, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":22.03,\"end\":22.44}]"} {"utterance_id": "4446-2273-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0001.flac", "task_type": "lexical_detection", "query": "hilda", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"hilda\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.91, 4.32, "hilda"]], "response_mode": "offline", "answer": "[{\"word\":\"hilda\",\"start\":3.91,\"end\":4.32}]"} {"utterance_id": "1580-141083-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0031.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.35, 0.74, "holmes"]], "response_mode": "offline", "answer": "[{\"word\":\"holmes\",\"start\":0.35,\"end\":0.74}]"} {"utterance_id": "5142-33396-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0019.flac", "task_type": "semantic_detection", "query": "food", "prompt": "After the audio ends, extract all occurrences of a food name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.37, 4.97, "meals"]], "response_mode": "offline", "answer": "[{\"word\":\"meals\",\"start\":4.37,\"end\":4.97}]"} {"utterance_id": "8224-274381-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0003.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.56, 3.41, "montrose's"]], "response_mode": "offline", "answer": "[{\"word\":\"montrose's\",\"start\":2.56,\"end\":3.41}]"} {"utterance_id": "237-134500-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0032.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.63, 2.13, "horses"]], "response_mode": "offline", "answer": "[{\"word\":\"horses\",\"start\":1.63,\"end\":2.13}]"} {"utterance_id": "7176-92135-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0027.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.2, 1.06, "lady larkspur"]], "response_mode": "offline", "answer": "[{\"word\":\"lady larkspur\",\"start\":0.2,\"end\":1.06}]"} {"utterance_id": "1320-122617-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0002.flac", "task_type": "lexical_detection", "query": "singing", "prompt": "After the audio ends, extract all occurrences of the word \"singing\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.71, 8.27, "singing"]], "response_mode": "offline", "answer": "[{\"word\":\"singing\",\"start\":7.71,\"end\":8.27}]"} {"utterance_id": "1221-135767-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0000.flac", "task_type": "lexical_detection", "query": "bellingham", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"bellingham\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.09, 3.9, "bellingham"]], "response_mode": "offline", "answer": "[{\"word\":\"bellingham\",\"start\":3.09,\"end\":3.9}]"} {"utterance_id": "4446-2271-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0016.flac", "task_type": "lexical_detection", "query": "slender", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"slender\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.09, 3.71, "slender"]], "response_mode": "offline", "answer": "[{\"word\":\"slender\",\"start\":3.09,\"end\":3.71}]"} {"utterance_id": "672-122797-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0039.flac", "task_type": "lexical_detection", "query": "fir", "prompt": "When the utterance is complete, report all instances of the word \"fir\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.16, 3.4, "fir"]], "response_mode": "offline", "answer": "[{\"word\":\"fir\",\"start\":3.16,\"end\":3.4}]"} {"utterance_id": "260-123288-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0025.flac", "task_type": "lexical_detection", "query": "leaps", "prompt": "Once the audio has finished, identify every instance of the word \"leaps\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.32, 6.77, "leaps"]], "response_mode": "offline", "answer": "[{\"word\":\"leaps\",\"start\":6.32,\"end\":6.77}]"} {"utterance_id": "6930-81414-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0024.flac", "task_type": "lexical_detection", "query": "articulate", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"articulate\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.41, 2.19, "articulate"]], "response_mode": "offline", "answer": "[{\"word\":\"articulate\",\"start\":1.41,\"end\":2.19}]"} {"utterance_id": "6930-76324-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0002.flac", "task_type": "lexical_detection", "query": "cynthia", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"cynthia\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.06, 2.51, "cynthia"]], "response_mode": "offline", "answer": "[{\"word\":\"cynthia\",\"start\":2.06,\"end\":2.51}]"} {"utterance_id": "121-127105-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0024.flac", "task_type": "semantic_counting", "query": "ordinal", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for an ordinal number with word/start/end. If none, reply [].", "labels": [[5.89, 6.18, "one"], [13.78, 14.24, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.89,\"end\":6.18},{\"word\":\"two\",\"start\":13.78,\"end\":14.24}]"} {"utterance_id": "7729-102255-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0032.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the full audio. After it finishes, find every instance of an ordinal number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.05, 13.44, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":13.05,\"end\":13.44}]"} {"utterance_id": "3575-170457-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0050.flac", "task_type": "lexical_detection", "query": "childhood", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"childhood\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.8, 2.42, "childhood"]], "response_mode": "offline", "answer": "[{\"word\":\"childhood\",\"start\":1.8,\"end\":2.42}]"} {"utterance_id": "2830-3980-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0052.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.88, 3.42, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":2.88,\"end\":3.42}]"} {"utterance_id": "7176-92135-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0043.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "Listen to the full audio. Count every occurrence of a profession. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.36, 7.78, "one"], [11.7, 11.99, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.36,\"end\":7.78},{\"word\":\"two\",\"start\":11.7,\"end\":11.99}]"} {"utterance_id": "1221-135767-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.3, 1.78, "pearl's"]], "response_mode": "offline", "answer": "[{\"word\":\"pearl's\",\"start\":1.3,\"end\":1.78}]"} {"utterance_id": "3570-5695-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0015.flac", "task_type": "lexical_detection", "query": "mobility", "prompt": "Once the audio has finished, identify every instance of the word \"mobility\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.22, 1.66, "mobility"]], "response_mode": "offline", "answer": "[{\"word\":\"mobility\",\"start\":1.22,\"end\":1.66}]"} {"utterance_id": "8463-294828-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0011.flac", "task_type": "lexical_detection", "query": "everywhere", "prompt": "When the utterance is complete, report all instances of the word \"everywhere\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.65, 2.16, "everywhere"]], "response_mode": "offline", "answer": "[{\"word\":\"everywhere\",\"start\":1.65,\"end\":2.16}]"} {"utterance_id": "4992-41806-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0015.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.54, 1.27, "missus harmon"]], "response_mode": "offline", "answer": "[{\"word\":\"missus harmon\",\"start\":0.54,\"end\":1.27}]"} {"utterance_id": "260-123286-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0000.flac", "task_type": "lexical_detection", "query": "saturday", "prompt": "When the utterance is complete, report all instances of the word \"saturday\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 1.1, "saturday"]], "response_mode": "offline", "answer": "[{\"word\":\"saturday\",\"start\":0.53,\"end\":1.1}]"} {"utterance_id": "8224-274381-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0013.flac", "task_type": "lexical_detection", "query": "companies", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"companies\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.15, 3.73, "companies"]], "response_mode": "offline", "answer": "[{\"word\":\"companies\",\"start\":3.15,\"end\":3.73}]"} {"utterance_id": "121-121726-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0010.flac", "task_type": "lexical_detection", "query": "domestic", "prompt": "Once the audio has finished, identify every instance of the word \"domestic\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.92, 3.61, "domestic"]], "response_mode": "offline", "answer": "[{\"word\":\"domestic\",\"start\":2.92,\"end\":3.61}]"} {"utterance_id": "1320-122617-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0041.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.26, 1.66, "skin"]], "response_mode": "offline", "answer": "[{\"word\":\"skin\",\"start\":1.26,\"end\":1.66}]"} {"utterance_id": "8455-210777-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0024.flac", "task_type": "lexical_counting", "query": "evil", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"evil\" with word/start/end. If none, reply [].", "labels": [[0.98, 1.55, "one"], [3.55, 4.01, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.98,\"end\":1.55},{\"word\":\"two\",\"start\":3.55,\"end\":4.01}]"} {"utterance_id": "7176-88083-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0024.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the full audio. After it finishes, find every instance of an ordinal number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.49, 0.96, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":0.49,\"end\":0.96}]"} {"utterance_id": "2830-3979-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0000.flac", "task_type": "lexical_detection", "query": "leading", "prompt": "When the utterance is complete, report all instances of the word \"leading\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.84, 2.1, "leading"]], "response_mode": "offline", "answer": "[{\"word\":\"leading\",\"start\":1.84,\"end\":2.1}]"} {"utterance_id": "6930-81414-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0002.flac", "task_type": "lexical_detection", "query": "onward", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"onward\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.41, 1.47, "onward"]], "response_mode": "offline", "answer": "[{\"word\":\"onward\",\"start\":0.41,\"end\":1.47}]"} {"utterance_id": "7127-75946-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0023.flac", "task_type": "lexical_detection", "query": "present", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"present\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.9, 3.49, "present"]], "response_mode": "offline", "answer": "[{\"word\":\"present\",\"start\":2.9,\"end\":3.49}]"} {"utterance_id": "5142-36377-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0007.flac", "task_type": "lexical_detection", "query": "cracked", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"cracked\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.55, 1.07, "cracked"]], "response_mode": "offline", "answer": "[{\"word\":\"cracked\",\"start\":0.55,\"end\":1.07}]"} {"utterance_id": "8224-274384-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0004.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[6.4, 6.87, "king"], [19.57, 20.14, "king"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":6.4,\"end\":6.87},{\"word\":\"king\",\"start\":19.57,\"end\":20.14}]"} {"utterance_id": "7127-75947-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0040.flac", "task_type": "lexical_counting", "query": "fete", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"fete\" (word, start, end). If none, reply [].", "labels": [[1.35, 1.6, "one"], [3.26, 3.57, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.35,\"end\":1.6},{\"word\":\"two\",\"start\":3.26,\"end\":3.57}]"} {"utterance_id": "2961-960-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0012.flac", "task_type": "lexical_detection", "query": "included", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"included\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.86, 5.57, "included"]], "response_mode": "offline", "answer": "[{\"word\":\"included\",\"start\":4.86,\"end\":5.57}]"} {"utterance_id": "2094-142345-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0036.flac", "task_type": "semantic_counting", "query": "kinship", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a family relation with word/start/end. If none, reply [].", "labels": [[1.02, 1.35, "one"], [5.75, 6.47, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.02,\"end\":1.35},{\"word\":\"two\",\"start\":5.75,\"end\":6.47}]"} {"utterance_id": "672-122797-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0050.flac", "task_type": "lexical_detection", "query": "branches", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"branches\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.68, 4.44, "branches"]], "response_mode": "offline", "answer": "[{\"word\":\"branches\",\"start\":3.68,\"end\":4.44}]"} {"utterance_id": "1221-135767-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0018.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.85, 1.33, "pearl"]], "response_mode": "offline", "answer": "[{\"word\":\"pearl\",\"start\":0.85,\"end\":1.33}]"} {"utterance_id": "3570-5696-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0003.flac", "task_type": "lexical_detection", "query": "resort", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"resort\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.98, 4.48, "resort"]], "response_mode": "offline", "answer": "[{\"word\":\"resort\",\"start\":3.98,\"end\":4.48}]"} {"utterance_id": "8455-210777-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0013.flac", "task_type": "lexical_detection", "query": "standing", "prompt": "Process the full audio offline. Find every instance of the word \"standing\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.16, 1.57, "standing"]], "response_mode": "offline", "answer": "[{\"word\":\"standing\",\"start\":1.16,\"end\":1.57}]"} {"utterance_id": "121-127105-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0018.flac", "task_type": "lexical_detection", "query": "departure", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"departure\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.38, 1.88, "departure"]], "response_mode": "offline", "answer": "[{\"word\":\"departure\",\"start\":1.38,\"end\":1.88}]"} {"utterance_id": "61-70968-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0041.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an ordinal number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.19, 2.61, "second"]], "response_mode": "offline", "answer": "[{\"word\":\"second\",\"start\":2.19,\"end\":2.61}]"} {"utterance_id": "4077-13751-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0005.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Do not answer until the audio ends. Then list each instance of an animal name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[7.54, 7.84, "birds"]], "response_mode": "offline", "answer": "[{\"word\":\"birds\",\"start\":7.54,\"end\":7.84}]"} {"utterance_id": "5142-33396-0060", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0060.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.71, 1.14, "thorkel"]], "response_mode": "offline", "answer": "[{\"word\":\"thorkel\",\"start\":0.71,\"end\":1.14}]"} {"utterance_id": "4992-41806-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0000.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen to the full audio. After it finishes, find every instance of a time of day. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.44, 4.86, "evening"]], "response_mode": "offline", "answer": "[{\"word\":\"evening\",\"start\":4.44,\"end\":4.86}]"} {"utterance_id": "3570-5696-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0007.flac", "task_type": "lexical_counting", "query": "waste", "prompt": "Listen to the entire clip, then report the count sequence for the word \"waste\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.96, 1.43, "one"], [8.79, 9.29, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.96,\"end\":1.43},{\"word\":\"two\",\"start\":8.79,\"end\":9.29}]"} {"utterance_id": "1284-1180-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0024.flac", "task_type": "lexical_detection", "query": "sympathetic", "prompt": "Process the full audio offline. Find every instance of the word \"sympathetic\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.99, 4.58, "sympathetic"]], "response_mode": "offline", "answer": "[{\"word\":\"sympathetic\",\"start\":3.99,\"end\":4.58}]"} {"utterance_id": "5142-36377-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0003.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a food name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.07, 3.44, "supper"]], "response_mode": "offline", "answer": "[{\"word\":\"supper\",\"start\":3.07,\"end\":3.44}]"} {"utterance_id": "8224-274384-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0010.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Once the audio has finished, identify every instance of a country name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.95, 7.67, "kingdoms"]], "response_mode": "offline", "answer": "[{\"word\":\"kingdoms\",\"start\":6.95,\"end\":7.67}]"} {"utterance_id": "8455-210777-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0004.flac", "task_type": "lexical_detection", "query": "distant", "prompt": "After listening to the whole recording, output every instance of the word \"distant\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.95, 6.3, "distant"]], "response_mode": "offline", "answer": "[{\"word\":\"distant\",\"start\":5.95,\"end\":6.3}]"} {"utterance_id": "1284-1180-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0002.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a color name with fields word, start, end (seconds). If none, reply [].", "labels": [[4.7, 4.97, "blue"], [6.42, 6.77, "gold"]], "response_mode": "offline", "answer": "[{\"word\":\"blue\",\"start\":4.7,\"end\":4.97},{\"word\":\"gold\",\"start\":6.42,\"end\":6.77}]"} {"utterance_id": "2300-131720-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0029.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.64, 1.09, "edison"]], "response_mode": "offline", "answer": "[{\"word\":\"edison\",\"start\":0.64,\"end\":1.09}]"} {"utterance_id": "1221-135767-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0013.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "After listening to the whole recording, output timed count words for a number (one, two, three, ...) as JSON. If none, reply [].", "labels": [[5.69, 5.9, "one"], [10.2, 10.6, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.69,\"end\":5.9},{\"word\":\"two\",\"start\":10.2,\"end\":10.6}]"} {"utterance_id": "8224-274381-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0014.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, list every instance of an organization name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.99, 2.57, "parliament"], [13.97, 14.44, "army"], [16.75, 17.39, "faction"]], "response_mode": "offline", "answer": "[{\"word\":\"parliament\",\"start\":1.99,\"end\":2.57},{\"word\":\"army\",\"start\":13.97,\"end\":14.44},{\"word\":\"faction\",\"start\":16.75,\"end\":17.39}]"} {"utterance_id": "5683-32879-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0003.flac", "task_type": "lexical_counting", "query": "better", "prompt": "After the audio ends, tally every instance of the word \"better\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[1.93, 2.26, "one"], [3.82, 4.11, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.93,\"end\":2.26},{\"word\":\"two\",\"start\":3.82,\"end\":4.11}]"} {"utterance_id": "3575-170457-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0037.flac", "task_type": "lexical_detection", "query": "perfection", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"perfection\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.99, 1.62, "perfection"]], "response_mode": "offline", "answer": "[{\"word\":\"perfection\",\"start\":0.99,\"end\":1.62}]"} {"utterance_id": "1580-141084-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0016.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.5, 3.94, "shoulders"]], "response_mode": "offline", "answer": "[{\"word\":\"shoulders\",\"start\":3.5,\"end\":3.94}]"} {"utterance_id": "2830-3980-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0020.flac", "task_type": "lexical_detection", "query": "sinful", "prompt": "When the utterance is complete, report all instances of the word \"sinful\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.83, 1.24, "sinful"]], "response_mode": "offline", "answer": "[{\"word\":\"sinful\",\"start\":0.83,\"end\":1.24}]"} {"utterance_id": "8463-294825-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0019.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Do not answer until the audio ends. Then list each instance of an ordinal number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.01, 2.5, "fifth"]], "response_mode": "offline", "answer": "[{\"word\":\"fifth\",\"start\":2.01,\"end\":2.5}]"} {"utterance_id": "7176-88083-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0021.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "You will hear a complete utterance. When it ends, list every instance of a weather term you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.19, 7.55, "cloud"]], "response_mode": "offline", "answer": "[{\"word\":\"cloud\",\"start\":7.19,\"end\":7.55}]"} {"utterance_id": "6829-68771-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0032.flac", "task_type": "lexical_detection", "query": "repress", "prompt": "After listening to the whole recording, output every instance of the word \"repress\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.08, 2.54, "repress"]], "response_mode": "offline", "answer": "[{\"word\":\"repress\",\"start\":2.08,\"end\":2.54}]"} {"utterance_id": "1221-135767-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0005.flac", "task_type": "lexical_detection", "query": "scarlet", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"scarlet\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.81, 1.27, "scarlet"], [3.28, 3.74, "scarlet"]], "response_mode": "offline", "answer": "[{\"word\":\"scarlet\",\"start\":0.81,\"end\":1.27},{\"word\":\"scarlet\",\"start\":3.28,\"end\":3.74}]"} {"utterance_id": "4970-29095-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0001.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Once the audio has finished, identify every instance of a time of day. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.33, 1.78, "morning"]], "response_mode": "offline", "answer": "[{\"word\":\"morning\",\"start\":1.33,\"end\":1.78}]"} {"utterance_id": "2830-3980-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0054.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.55, 0.94, "christ"], [3.19, 3.46, "paul"]], "response_mode": "offline", "answer": "[{\"word\":\"christ\",\"start\":0.55,\"end\":0.94},{\"word\":\"paul\",\"start\":3.19,\"end\":3.46}]"} {"utterance_id": "7127-75947-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0008.flac", "task_type": "lexical_detection", "query": "heart", "prompt": "When the utterance is complete, report all instances of the word \"heart\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.52, 1.93, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":1.52,\"end\":1.93}]"} {"utterance_id": "8463-294828-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0000.flac", "task_type": "lexical_detection", "query": "wishes", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"wishes\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.37, 3.08, "wishes"]], "response_mode": "offline", "answer": "[{\"word\":\"wishes\",\"start\":2.37,\"end\":3.08}]"} {"utterance_id": "237-134500-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0041.flac", "task_type": "lexical_detection", "query": "slowly", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"slowly\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.79, 4.3, "slowly"]], "response_mode": "offline", "answer": "[{\"word\":\"slowly\",\"start\":3.79,\"end\":4.3}]"} {"utterance_id": "5105-28240-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0018.flac", "task_type": "lexical_detection", "query": "board", "prompt": "When the utterance is complete, report all instances of the word \"board\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.13, 1.45, "board"]], "response_mode": "offline", "answer": "[{\"word\":\"board\",\"start\":1.13,\"end\":1.45}]"} {"utterance_id": "8555-284449-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0006.flac", "task_type": "lexical_detection", "query": "pretty", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"pretty\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.02, 3.29, "pretty"]], "response_mode": "offline", "answer": "[{\"word\":\"pretty\",\"start\":3.02,\"end\":3.29}]"} {"utterance_id": "7021-85628-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0008.flac", "task_type": "lexical_detection", "query": "cap", "prompt": "When the utterance is complete, report all instances of the word \"cap\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.32, 6.81, "cap"]], "response_mode": "offline", "answer": "[{\"word\":\"cap\",\"start\":6.32,\"end\":6.81}]"} {"utterance_id": "1221-135766-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0011.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a person's name (word, start, end). If none, reply [].", "labels": [[1.28, 1.61, "one"], [17.45, 17.85, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.28,\"end\":1.61},{\"word\":\"two\",\"start\":17.45,\"end\":17.85}]"} {"utterance_id": "7021-79730-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.76, 5.2, "mary"]], "response_mode": "offline", "answer": "[{\"word\":\"mary\",\"start\":4.76,\"end\":5.2}]"} {"utterance_id": "3729-6852-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0017.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.19, 3.59, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.19,\"end\":3.59}]"} {"utterance_id": "5142-33396-0062", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0062.flac", "task_type": "lexical_detection", "query": "arm", "prompt": "When the utterance is complete, report all instances of the word \"arm\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.31, 1.54, "arm"]], "response_mode": "offline", "answer": "[{\"word\":\"arm\",\"start\":1.31,\"end\":1.54}]"} {"utterance_id": "2094-142345-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0033.flac", "task_type": "lexical_detection", "query": "eleven", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"eleven\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.77, 3.07, "eleven"]], "response_mode": "offline", "answer": "[{\"word\":\"eleven\",\"start\":2.77,\"end\":3.07}]"} {"utterance_id": "1580-141083-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0006.flac", "task_type": "lexical_detection", "query": "belonged", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"belonged\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.28, 3.62, "belonged"]], "response_mode": "offline", "answer": "[{\"word\":\"belonged\",\"start\":3.28,\"end\":3.62}]"} {"utterance_id": "6829-68769-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0027.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Do not answer until the audio ends. Then list each instance of a money or currency term as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.52, 1.97, "check"]], "response_mode": "offline", "answer": "[{\"word\":\"check\",\"start\":1.52,\"end\":1.97}]"} {"utterance_id": "3729-6852-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0001.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.96, 11.73, "insect"]], "response_mode": "offline", "answer": "[{\"word\":\"insect\",\"start\":10.96,\"end\":11.73}]"} {"utterance_id": "1089-134686-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0023.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.12, 7.06, "jesus christ"]], "response_mode": "offline", "answer": "[{\"word\":\"jesus christ\",\"start\":6.12,\"end\":7.06}]"} {"utterance_id": "6930-75918-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0018.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Listen carefully to the entire clip, then report every instance of a weather term as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.35, 3.81, "tempest"]], "response_mode": "offline", "answer": "[{\"word\":\"tempest\",\"start\":3.35,\"end\":3.81}]"} {"utterance_id": "1221-135766-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0011.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.55, 11.02, "bosom"], [18.1, 18.4, "flesh"]], "response_mode": "offline", "answer": "[{\"word\":\"bosom\",\"start\":10.55,\"end\":11.02},{\"word\":\"flesh\",\"start\":18.1,\"end\":18.4}]"} {"utterance_id": "121-121726-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0007.flac", "task_type": "lexical_detection", "query": "wisdom", "prompt": "Once the audio has finished, identify every instance of the word \"wisdom\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.38, 3.97, "wisdom"]], "response_mode": "offline", "answer": "[{\"word\":\"wisdom\",\"start\":3.38,\"end\":3.97}]"} {"utterance_id": "4077-13751-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0018.flac", "task_type": "semantic_counting", "query": "organization", "prompt": "Once the audio has finished, list each occurrence of an organization name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[0.99, 1.31, "one"], [4.43, 5.37, "two"], [11.51, 12.13, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.99,\"end\":1.31},{\"word\":\"two\",\"start\":4.43,\"end\":5.37},{\"word\":\"three\",\"start\":11.51,\"end\":12.13}]"} {"utterance_id": "3575-170457-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0020.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.6, 0.9, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":0.6,\"end\":0.9}]"} {"utterance_id": "6829-68769-0049", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0049.flac", "task_type": "lexical_detection", "query": "grasping", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"grasping\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.0, 1.55, "grasping"]], "response_mode": "offline", "answer": "[{\"word\":\"grasping\",\"start\":1.0,\"end\":1.55}]"} {"utterance_id": "4446-2273-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0034.flac", "task_type": "lexical_detection", "query": "figure", "prompt": "After listening to the whole recording, output every instance of the word \"figure\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.17, 2.6, "figure"]], "response_mode": "offline", "answer": "[{\"word\":\"figure\",\"start\":2.17,\"end\":2.6}]"} {"utterance_id": "3575-170457-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0008.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the full audio. After it finishes, find every instance of a family relation. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.36, 9.69, "daughter"]], "response_mode": "offline", "answer": "[{\"word\":\"daughter\",\"start\":9.36,\"end\":9.69}]"} {"utterance_id": "8555-284447-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0002.flac", "task_type": "semantic_detection", "query": "food", "prompt": "When the utterance is complete, report all instances of a food name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.37, 1.68, "coffee"], [7.14, 7.81, "breakfas"]], "response_mode": "offline", "answer": "[{\"word\":\"coffee\",\"start\":1.37,\"end\":1.68},{\"word\":\"breakfas\",\"start\":7.14,\"end\":7.81}]"} {"utterance_id": "237-134500-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0010.flac", "task_type": "lexical_detection", "query": "exciting", "prompt": "Once the audio has finished, identify every instance of the word \"exciting\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.34, 0.8, "exciting"]], "response_mode": "offline", "answer": "[{\"word\":\"exciting\",\"start\":0.34,\"end\":0.8}]"} {"utterance_id": "61-70968-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0010.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.55, 5.98, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":5.55,\"end\":5.98}]"} {"utterance_id": "5683-32879-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.18, 2.7, "rachel"]], "response_mode": "offline", "answer": "[{\"word\":\"rachel\",\"start\":2.18,\"end\":2.7}]"} {"utterance_id": "2094-142345-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0028.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[2.72, 3.02, "feet"]], "response_mode": "offline", "answer": "[{\"word\":\"feet\",\"start\":2.72,\"end\":3.02}]"} {"utterance_id": "1284-1181-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0011.flac", "task_type": "lexical_counting", "query": "golden", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"golden\" (word, start, end). If none, reply [].", "labels": [[4.6, 5.06, "one"], [6.59, 6.97, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.6,\"end\":5.06},{\"word\":\"two\",\"start\":6.59,\"end\":6.97}]"} {"utterance_id": "5639-40744-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0034.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.9, 10.12, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":9.9,\"end\":10.12}]"} {"utterance_id": "6930-75918-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.76, 2.16, "raoul"]], "response_mode": "offline", "answer": "[{\"word\":\"raoul\",\"start\":1.76,\"end\":2.16}]"} {"utterance_id": "5105-28241-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0011.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a country name with fields word, start, end (seconds). If none, reply [].", "labels": [[5.94, 6.48, "france"]], "response_mode": "offline", "answer": "[{\"word\":\"france\",\"start\":5.94,\"end\":6.48}]"} {"utterance_id": "4970-29093-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0019.flac", "task_type": "lexical_detection", "query": "spent", "prompt": "After listening to the whole recording, output every instance of the word \"spent\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.04, 1.4, "spent"]], "response_mode": "offline", "answer": "[{\"word\":\"spent\",\"start\":1.04,\"end\":1.4}]"} {"utterance_id": "6829-68769-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0019.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the full audio. After it finishes, find every instance of an organization name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.19, 2.65, "jail"]], "response_mode": "offline", "answer": "[{\"word\":\"jail\",\"start\":2.19,\"end\":2.65}]"} {"utterance_id": "5683-32879-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.42, 6.74, "rachel"]], "response_mode": "offline", "answer": "[{\"word\":\"rachel\",\"start\":6.42,\"end\":6.74}]"} {"utterance_id": "4446-2271-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0008.flac", "task_type": "lexical_detection", "query": "somewhere", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"somewhere\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.02, 4.34, "somewhere"]], "response_mode": "offline", "answer": "[{\"word\":\"somewhere\",\"start\":4.02,\"end\":4.34}]"} {"utterance_id": "1580-141083-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0043.flac", "task_type": "lexical_detection", "query": "miles", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"miles\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.68, 1.98, "miles"]], "response_mode": "offline", "answer": "[{\"word\":\"miles\",\"start\":1.68,\"end\":1.98}]"} {"utterance_id": "4446-2273-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0011.flac", "task_type": "lexical_detection", "query": "looks", "prompt": "When the utterance is complete, report all instances of the word \"looks\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.45, 1.7, "looks"]], "response_mode": "offline", "answer": "[{\"word\":\"looks\",\"start\":1.45,\"end\":1.7}]"} {"utterance_id": "7729-102255-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0038.flac", "task_type": "lexical_detection", "query": "mob", "prompt": "After the audio ends, extract all occurrences of the word \"mob\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.66, 3.08, "mob"]], "response_mode": "offline", "answer": "[{\"word\":\"mob\",\"start\":2.66,\"end\":3.08}]"} {"utterance_id": "1221-135766-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0008.flac", "task_type": "lexical_detection", "query": "strict", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"strict\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.23, 6.6, "strict"]], "response_mode": "offline", "answer": "[{\"word\":\"strict\",\"start\":6.23,\"end\":6.6}]"} {"utterance_id": "1089-134686-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0018.flac", "task_type": "lexical_detection", "query": "condemnation", "prompt": "After the audio ends, extract all occurrences of the word \"condemnation\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.25, 15.17, "condemnation"]], "response_mode": "offline", "answer": "[{\"word\":\"condemnation\",\"start\":14.25,\"end\":15.17}]"} {"utterance_id": "6829-68769-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0010.flac", "task_type": "lexical_counting", "query": "talk", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"talk\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.82, 1.11, "one"], [2.67, 3.12, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.82,\"end\":1.11},{\"word\":\"two\",\"start\":2.67,\"end\":3.12}]"} {"utterance_id": "5142-33396-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0036.flac", "task_type": "lexical_detection", "query": "law", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"law\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.64, 2.07, "law"]], "response_mode": "offline", "answer": "[{\"word\":\"law\",\"start\":1.64,\"end\":2.07}]"} {"utterance_id": "121-121726-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0003.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.79, 3.16, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":2.79,\"end\":3.16}]"} {"utterance_id": "8463-294828-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0001.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.49, 3.6, "j b hobson's"]], "response_mode": "offline", "answer": "[{\"word\":\"j b hobson's\",\"start\":2.49,\"end\":3.6}]"} {"utterance_id": "1995-1837-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0008.flac", "task_type": "lexical_detection", "query": "use", "prompt": "After listening to the whole recording, output every instance of the word \"use\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.77, 1.01, "use"]], "response_mode": "offline", "answer": "[{\"word\":\"use\",\"start\":0.77,\"end\":1.01}]"} {"utterance_id": "1221-135767-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0023.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Once the audio has finished, identify every instance of a family relation. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.59, 5.28, "descendants"]], "response_mode": "offline", "answer": "[{\"word\":\"descendants\",\"start\":4.59,\"end\":5.28}]"} {"utterance_id": "6930-76324-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0014.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.39, 2.71, "heads"]], "response_mode": "offline", "answer": "[{\"word\":\"heads\",\"start\":2.39,\"end\":2.71}]"} {"utterance_id": "1188-133604-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0034.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.06, 13.54, "chaucer"]], "response_mode": "offline", "answer": "[{\"word\":\"chaucer\",\"start\":13.06,\"end\":13.54}]"} {"utterance_id": "61-70968-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0053.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.77, 3.06, "robin"]], "response_mode": "offline", "answer": "[{\"word\":\"robin\",\"start\":2.77,\"end\":3.06}]"} {"utterance_id": "3575-170457-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0036.flac", "task_type": "lexical_detection", "query": "fill", "prompt": "After the audio ends, extract all occurrences of the word \"fill\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.05, 1.2, "fill"]], "response_mode": "offline", "answer": "[{\"word\":\"fill\",\"start\":1.05,\"end\":1.2}]"} {"utterance_id": "4970-29093-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0000.flac", "task_type": "lexical_detection", "query": "astor", "prompt": "After the audio ends, extract all occurrences of the word \"astor\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.46, 1.83, "astor"]], "response_mode": "offline", "answer": "[{\"word\":\"astor\",\"start\":1.46,\"end\":1.83}]"} {"utterance_id": "1089-134691-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0020.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.01, 1.94, "stephanos"], [3.25, 3.86, "dedalus"]], "response_mode": "offline", "answer": "[{\"word\":\"stephanos\",\"start\":1.01,\"end\":1.94},{\"word\":\"dedalus\",\"start\":3.25,\"end\":3.86}]"} {"utterance_id": "1221-135767-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0006.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.05, 5.58, "brain"]], "response_mode": "offline", "answer": "[{\"word\":\"brain\",\"start\":5.05,\"end\":5.58}]"} {"utterance_id": "7127-75946-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0026.flac", "task_type": "lexical_detection", "query": "culminating", "prompt": "Process the full audio offline. Find every instance of the word \"culminating\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.61, 15.23, "culminating"]], "response_mode": "offline", "answer": "[{\"word\":\"culminating\",\"start\":14.61,\"end\":15.23}]"} {"utterance_id": "8463-294828-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0005.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.23, 0.89, "conseil"]], "response_mode": "offline", "answer": "[{\"word\":\"conseil\",\"start\":0.23,\"end\":0.89}]"} {"utterance_id": "672-122797-0071", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0071.flac", "task_type": "lexical_detection", "query": "round", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"round\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.14, 5.49, "round"]], "response_mode": "offline", "answer": "[{\"word\":\"round\",\"start\":5.14,\"end\":5.49}]"} {"utterance_id": "5142-33396-0061", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0061.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a family relation with fields word, start, end (seconds). If none, reply [].", "labels": [[0.29, 0.66, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":0.29,\"end\":0.66}]"} {"utterance_id": "1089-134686-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0029.flac", "task_type": "lexical_detection", "query": "afternoon", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"afternoon\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.97, 3.58, "afternoon"]], "response_mode": "offline", "answer": "[{\"word\":\"afternoon\",\"start\":2.97,\"end\":3.58}]"} {"utterance_id": "121-127105-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0002.flac", "task_type": "lexical_detection", "query": "notice", "prompt": "After the audio ends, extract all occurrences of the word \"notice\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.94, 2.31, "notice"]], "response_mode": "offline", "answer": "[{\"word\":\"notice\",\"start\":1.94,\"end\":2.31}]"} {"utterance_id": "1221-135767-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0010.flac", "task_type": "lexical_detection", "query": "doubtless", "prompt": "Process the full audio offline. Find every instance of the word \"doubtless\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.4, 4.87, "doubtless"]], "response_mode": "offline", "answer": "[{\"word\":\"doubtless\",\"start\":4.4,\"end\":4.87}]"} {"utterance_id": "7021-79730-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0004.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.22, 7.96, "countenance"], [13.44, 13.95, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"countenance\",\"start\":7.22,\"end\":7.96},{\"word\":\"face\",\"start\":13.44,\"end\":13.95}]"} {"utterance_id": "2830-3980-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0022.flac", "task_type": "lexical_detection", "query": "sight", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"sight\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.75, 3.1, "sight"]], "response_mode": "offline", "answer": "[{\"word\":\"sight\",\"start\":2.75,\"end\":3.1}]"} {"utterance_id": "7176-92135-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0041.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.29, 1.65, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":1.29,\"end\":1.65}]"} {"utterance_id": "1284-1180-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0014.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 0.92, "ojo"]], "response_mode": "offline", "answer": "[{\"word\":\"ojo\",\"start\":0.51,\"end\":0.92}]"} {"utterance_id": "7127-75946-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0000.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, list every instance of an organization name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[18.73, 19.14, "court"]], "response_mode": "offline", "answer": "[{\"word\":\"court\",\"start\":18.73,\"end\":19.14}]"} {"utterance_id": "1284-1180-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0005.flac", "task_type": "lexical_detection", "query": "disturb", "prompt": "Once the audio has finished, identify every instance of the word \"disturb\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.82, 1.28, "disturb"]], "response_mode": "offline", "answer": "[{\"word\":\"disturb\",\"start\":0.82,\"end\":1.28}]"} {"utterance_id": "2300-131720-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0019.flac", "task_type": "lexical_detection", "query": "enormously", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"enormously\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.54, 8.2, "enormously"]], "response_mode": "offline", "answer": "[{\"word\":\"enormously\",\"start\":7.54,\"end\":8.2}]"} {"utterance_id": "5142-33396-0061", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0061.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.77, 1.08, "queen"]], "response_mode": "offline", "answer": "[{\"word\":\"queen\",\"start\":0.77,\"end\":1.08}]"} {"utterance_id": "4970-29095-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0028.flac", "task_type": "lexical_detection", "query": "marked", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"marked\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.85, 6.32, "marked"]], "response_mode": "offline", "answer": "[{\"word\":\"marked\",\"start\":5.85,\"end\":6.32}]"} {"utterance_id": "7127-75946-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0002.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[11.83, 12.05, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":11.83,\"end\":12.05}]"} {"utterance_id": "4992-41806-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0008.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.27, 6.64, "next"]], "response_mode": "offline", "answer": "[{\"word\":\"next\",\"start\":6.27,\"end\":6.64}]"} {"utterance_id": "1284-1180-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0010.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the full audio. After it finishes, find every instance of a color name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.87, 5.34, "blue"]], "response_mode": "offline", "answer": "[{\"word\":\"blue\",\"start\":4.87,\"end\":5.34}]"} {"utterance_id": "5683-32865-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0007.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.22, 5.74, "shoulder"]], "response_mode": "offline", "answer": "[{\"word\":\"shoulder\",\"start\":5.22,\"end\":5.74}]"} {"utterance_id": "1221-135767-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0023.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[15.07, 15.34, "back"]], "response_mode": "offline", "answer": "[{\"word\":\"back\",\"start\":15.07,\"end\":15.34}]"} {"utterance_id": "1089-134691-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0010.flac", "task_type": "lexical_counting", "query": "brother", "prompt": "After the audio ends, tally every instance of the word \"brother\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.4, 0.64, "one"], [2.08, 2.35, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.4,\"end\":0.64},{\"word\":\"two\",\"start\":2.08,\"end\":2.35}]"} {"utterance_id": "2300-131720-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0017.flac", "task_type": "semantic_detection", "query": "city", "prompt": "When the utterance is complete, report all instances of a city name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.75, 5.24, "new york"]], "response_mode": "offline", "answer": "[{\"word\":\"new york\",\"start\":4.75,\"end\":5.24}]"} {"utterance_id": "61-70970-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0029.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.95, 3.67, "warrenton"]], "response_mode": "offline", "answer": "[{\"word\":\"warrenton\",\"start\":2.95,\"end\":3.67}]"} {"utterance_id": "237-134493-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0009.flac", "task_type": "lexical_detection", "query": "emil", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"emil\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.9, 2.23, "emil"]], "response_mode": "offline", "answer": "[{\"word\":\"emil\",\"start\":1.9,\"end\":2.23}]"} {"utterance_id": "8230-279154-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0010.flac", "task_type": "lexical_detection", "query": "example", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"example\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[7.23, 7.84, "example"]], "response_mode": "offline", "answer": "[{\"word\":\"example\",\"start\":7.23,\"end\":7.84}]"} {"utterance_id": "7176-88083-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0009.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.54, 0.88, "hawk"]], "response_mode": "offline", "answer": "[{\"word\":\"hawk\",\"start\":0.54,\"end\":0.88}]"} {"utterance_id": "5142-33396-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0019.flac", "task_type": "lexical_detection", "query": "meals", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"meals\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.37, 4.97, "meals"]], "response_mode": "offline", "answer": "[{\"word\":\"meals\",\"start\":4.37,\"end\":4.97}]"} {"utterance_id": "1284-1180-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0031.flac", "task_type": "lexical_detection", "query": "lives", "prompt": "After the audio ends, extract all occurrences of the word \"lives\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.7, 3.05, "lives"]], "response_mode": "offline", "answer": "[{\"word\":\"lives\",\"start\":2.7,\"end\":3.05}]"} {"utterance_id": "237-134500-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0031.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.82, 7.14, "arm"]], "response_mode": "offline", "answer": "[{\"word\":\"arm\",\"start\":6.82,\"end\":7.14}]"} {"utterance_id": "4970-29095-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0033.flac", "task_type": "lexical_detection", "query": "shut", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"shut\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[10.09, 10.36, "shut"]], "response_mode": "offline", "answer": "[{\"word\":\"shut\",\"start\":10.09,\"end\":10.36}]"} {"utterance_id": "2300-131720-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0024.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.31, 2.67, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":2.31,\"end\":2.67}]"} {"utterance_id": "237-134500-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0001.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 0.94, "marie"]], "response_mode": "offline", "answer": "[{\"word\":\"marie\",\"start\":0.51,\"end\":0.94}]"} {"utterance_id": "3570-5694-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0013.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a money or currency term with fields word, start, end (seconds). If none, reply [].", "labels": [[2.8, 3.15, "wealth"]], "response_mode": "offline", "answer": "[{\"word\":\"wealth\",\"start\":2.8,\"end\":3.15}]"} {"utterance_id": "4992-41806-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.91, "olive"], [5.8, 6.7, "mister popham"]], "response_mode": "offline", "answer": "[{\"word\":\"olive\",\"start\":0.53,\"end\":0.91},{\"word\":\"mister popham\",\"start\":5.8,\"end\":6.7}]"} {"utterance_id": "1284-1180-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0015.flac", "task_type": "lexical_detection", "query": "refresh", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"refresh\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.59, 5.04, "refresh"]], "response_mode": "offline", "answer": "[{\"word\":\"refresh\",\"start\":4.59,\"end\":5.04}]"} {"utterance_id": "8455-210777-0061", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0061.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After listening to the whole recording, output every instance of an organization name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.22, 4.19, "north north west birmingham regiment"]], "response_mode": "offline", "answer": "[{\"word\":\"north north west birmingham regiment\",\"start\":2.22,\"end\":4.19}]"} {"utterance_id": "121-123859-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123859/121-123859-0003.flac", "task_type": "lexical_detection", "query": "grow", "prompt": "Once the audio has finished, identify every instance of the word \"grow\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[10.02, 10.41, "grow"]], "response_mode": "offline", "answer": "[{\"word\":\"grow\",\"start\":10.02,\"end\":10.41}]"} {"utterance_id": "2961-960-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0021.flac", "task_type": "lexical_detection", "query": "flow", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"flow\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.75, 2.37, "flow"]], "response_mode": "offline", "answer": "[{\"word\":\"flow\",\"start\":1.75,\"end\":2.37}]"} {"utterance_id": "8455-210777-0066", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0066.flac", "task_type": "lexical_detection", "query": "course", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"course\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.76, 1.27, "course"]], "response_mode": "offline", "answer": "[{\"word\":\"course\",\"start\":0.76,\"end\":1.27}]"} {"utterance_id": "1221-135767-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0014.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a profession: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[3.66, 4.25, "one"], [6.15, 6.74, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.66,\"end\":4.25},{\"word\":\"two\",\"start\":6.15,\"end\":6.74}]"} {"utterance_id": "4992-23283-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0016.flac", "task_type": "lexical_detection", "query": "searched", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"searched\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.81, 1.2, "searched"]], "response_mode": "offline", "answer": "[{\"word\":\"searched\",\"start\":0.81,\"end\":1.2}]"} {"utterance_id": "3729-6852-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0044.flac", "task_type": "lexical_detection", "query": "insatiable", "prompt": "Once the audio has finished, identify every instance of the word \"insatiable\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.63, 6.48, "insatiable"]], "response_mode": "offline", "answer": "[{\"word\":\"insatiable\",\"start\":5.63,\"end\":6.48}]"} {"utterance_id": "3729-6852-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0005.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[8.3, 8.79, "authors"]], "response_mode": "offline", "answer": "[{\"word\":\"authors\",\"start\":8.3,\"end\":8.79}]"} {"utterance_id": "3575-170457-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0054.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a person's name (word, start, end). If none, reply [].", "labels": [[4.53, 5.5, "one"], [6.95, 7.22, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.53,\"end\":5.5},{\"word\":\"two\",\"start\":6.95,\"end\":7.22}]"} {"utterance_id": "4970-29095-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0031.flac", "task_type": "semantic_detection", "query": "country", "prompt": "You will hear a complete utterance. When it ends, list every instance of a country name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.98, 14.58, "country"]], "response_mode": "offline", "answer": "[{\"word\":\"country\",\"start\":13.98,\"end\":14.58}]"} {"utterance_id": "4446-2273-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0035.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.11, 1.48, "shoulder"], [2.94, 3.28, "ear"]], "response_mode": "offline", "answer": "[{\"word\":\"shoulder\",\"start\":1.11,\"end\":1.48},{\"word\":\"ear\",\"start\":2.94,\"end\":3.28}]"} {"utterance_id": "5639-40744-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0004.flac", "task_type": "lexical_detection", "query": "cloaks", "prompt": "Once the audio has finished, identify every instance of the word \"cloaks\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.58, 4.16, "cloaks"]], "response_mode": "offline", "answer": "[{\"word\":\"cloaks\",\"start\":3.58,\"end\":4.16}]"} {"utterance_id": "61-70968-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0041.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.03, 1.38, "will"], [2.61, 2.81, "will"]], "response_mode": "offline", "answer": "[{\"word\":\"will\",\"start\":1.03,\"end\":1.38},{\"word\":\"will\",\"start\":2.61,\"end\":2.81}]"} {"utterance_id": "1580-141084-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0028.flac", "task_type": "lexical_detection", "query": "man", "prompt": "Once the audio has finished, identify every instance of the word \"man\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.19, 1.68, "man"]], "response_mode": "offline", "answer": "[{\"word\":\"man\",\"start\":1.19,\"end\":1.68}]"} {"utterance_id": "908-31957-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0006.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.88, 1.24, "heart"], [4.11, 4.57, "wings"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":0.88,\"end\":1.24},{\"word\":\"wings\",\"start\":4.11,\"end\":4.57}]"} {"utterance_id": "1284-134647-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0006.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.26, 0.92, "bishops"]], "response_mode": "offline", "answer": "[{\"word\":\"bishops\",\"start\":0.26,\"end\":0.92}]"} {"utterance_id": "61-70970-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0024.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.39, 5.83, "warrenton"]], "response_mode": "offline", "answer": "[{\"word\":\"warrenton\",\"start\":5.39,\"end\":5.83}]"} {"utterance_id": "1221-135767-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0001.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Do not answer until the audio ends. Then list each instance of a city name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[12.42, 13.05, "settlement"]], "response_mode": "offline", "answer": "[{\"word\":\"settlement\",\"start\":12.42,\"end\":13.05}]"} {"utterance_id": "8455-210777-0063", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0063.flac", "task_type": "lexical_detection", "query": "shouted", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"shouted\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.26, 1.78, "shouted"]], "response_mode": "offline", "answer": "[{\"word\":\"shouted\",\"start\":1.26,\"end\":1.78}]"} {"utterance_id": "237-134493-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0005.flac", "task_type": "lexical_detection", "query": "straight", "prompt": "Process the full audio offline. Find every instance of the word \"straight\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.38, 3.87, "straight"]], "response_mode": "offline", "answer": "[{\"word\":\"straight\",\"start\":3.38,\"end\":3.87}]"} {"utterance_id": "8230-279154-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0037.flac", "task_type": "lexical_detection", "query": "associated", "prompt": "After the audio ends, extract all occurrences of the word \"associated\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.96, 10.83, "associated"]], "response_mode": "offline", "answer": "[{\"word\":\"associated\",\"start\":9.96,\"end\":10.83}]"} {"utterance_id": "5683-32879-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0012.flac", "task_type": "lexical_detection", "query": "cousin", "prompt": "Once the audio has finished, identify every instance of the word \"cousin\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.92, 2.44, "cousin"]], "response_mode": "offline", "answer": "[{\"word\":\"cousin\",\"start\":1.92,\"end\":2.44}]"} {"utterance_id": "6829-68771-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0015.flac", "task_type": "lexical_detection", "query": "martha", "prompt": "After listening to the whole recording, output every instance of the word \"martha\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.1, 2.39, "martha"]], "response_mode": "offline", "answer": "[{\"word\":\"martha\",\"start\":2.1,\"end\":2.39}]"} {"utterance_id": "2094-142345-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0033.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.83, 7.03, "aunt"]], "response_mode": "offline", "answer": "[{\"word\":\"aunt\",\"start\":6.83,\"end\":7.03}]"} {"utterance_id": "3575-170457-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0043.flac", "task_type": "lexical_detection", "query": "namely", "prompt": "After the audio ends, extract all occurrences of the word \"namely\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.55, 4.04, "namely"]], "response_mode": "offline", "answer": "[{\"word\":\"namely\",\"start\":3.55,\"end\":4.04}]"} {"utterance_id": "7021-85628-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0014.flac", "task_type": "lexical_detection", "query": "shook", "prompt": "Once the audio has finished, identify every instance of the word \"shook\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.2, 1.47, "shook"]], "response_mode": "offline", "answer": "[{\"word\":\"shook\",\"start\":1.2,\"end\":1.47}]"} {"utterance_id": "1089-134686-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0022.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.0, 5.41, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":5.0,\"end\":5.41}]"} {"utterance_id": "5105-28240-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0020.flac", "task_type": "lexical_detection", "query": "shook", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"shook\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.09, 1.38, "shook"]], "response_mode": "offline", "answer": "[{\"word\":\"shook\",\"start\":1.09,\"end\":1.38}]"} {"utterance_id": "237-134500-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0009.flac", "task_type": "lexical_detection", "query": "wet", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"wet\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.97, 1.18, "wet"]], "response_mode": "offline", "answer": "[{\"word\":\"wet\",\"start\":0.97,\"end\":1.18}]"} {"utterance_id": "4077-13754-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0010.flac", "task_type": "lexical_detection", "query": "national", "prompt": "Process the full audio offline. Find every instance of the word \"national\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.17, 7.63, "national"]], "response_mode": "offline", "answer": "[{\"word\":\"national\",\"start\":7.17,\"end\":7.63}]"} {"utterance_id": "3729-6852-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.51, 5.06, "silvia"], [16.6, 17.17, "silvia"]], "response_mode": "offline", "answer": "[{\"word\":\"silvia\",\"start\":4.51,\"end\":5.06},{\"word\":\"silvia\",\"start\":16.6,\"end\":17.17}]"} {"utterance_id": "1221-135766-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.53, 1.44, "hester prynne"]], "response_mode": "offline", "answer": "[{\"word\":\"hester prynne\",\"start\":0.53,\"end\":1.44}]"} {"utterance_id": "237-134500-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0030.flac", "task_type": "lexical_detection", "query": "account", "prompt": "Once the audio has finished, identify every instance of the word \"account\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.77, 2.27, "account"]], "response_mode": "offline", "answer": "[{\"word\":\"account\",\"start\":1.77,\"end\":2.27}]"} {"utterance_id": "4970-29095-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0030.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.03, 2.68, "philip"]], "response_mode": "offline", "answer": "[{\"word\":\"philip\",\"start\":2.03,\"end\":2.68}]"} {"utterance_id": "4992-23283-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0007.flac", "task_type": "lexical_detection", "query": "questions", "prompt": "When the utterance is complete, report all instances of the word \"questions\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.41, 1.98, "questions"]], "response_mode": "offline", "answer": "[{\"word\":\"questions\",\"start\":1.41,\"end\":1.98}]"} {"utterance_id": "908-157963-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0019.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.98, 1.71, "tenfold"]], "response_mode": "offline", "answer": "[{\"word\":\"tenfold\",\"start\":0.98,\"end\":1.71}]"} {"utterance_id": "2094-142345-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0011.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.92, 13.37, "hair"]], "response_mode": "offline", "answer": "[{\"word\":\"hair\",\"start\":12.92,\"end\":13.37}]"} {"utterance_id": "61-70968-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0010.flac", "task_type": "lexical_detection", "query": "peeped", "prompt": "When the utterance is complete, report all instances of the word \"peeped\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.04, 5.39, "peeped"]], "response_mode": "offline", "answer": "[{\"word\":\"peeped\",\"start\":5.04,\"end\":5.39}]"} {"utterance_id": "4970-29095-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0022.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.33, 2.83, "ruth"]], "response_mode": "offline", "answer": "[{\"word\":\"ruth\",\"start\":2.33,\"end\":2.83}]"} {"utterance_id": "8455-210777-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0032.flac", "task_type": "lexical_detection", "query": "duty", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"duty\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.9, 1.31, "duty"]], "response_mode": "offline", "answer": "[{\"word\":\"duty\",\"start\":0.9,\"end\":1.31}]"} {"utterance_id": "7176-88083-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0022.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an animal name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.34, 0.72, "hawk"], [5.43, 5.88, "weasel"]], "response_mode": "offline", "answer": "[{\"word\":\"hawk\",\"start\":0.34,\"end\":0.72},{\"word\":\"weasel\",\"start\":5.43,\"end\":5.88}]"} {"utterance_id": "8555-284449-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0009.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.25, 2.76, "sailor"]], "response_mode": "offline", "answer": "[{\"word\":\"sailor\",\"start\":2.25,\"end\":2.76}]"} {"utterance_id": "1320-122617-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.48, 3.09, "hawkeye"]], "response_mode": "offline", "answer": "[{\"word\":\"hawkeye\",\"start\":2.48,\"end\":3.09}]"} {"utterance_id": "7127-75947-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0028.flac", "task_type": "lexical_detection", "query": "montalais", "prompt": "Once the audio has finished, identify every instance of the word \"montalais\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.07, 3.7, "montalais"]], "response_mode": "offline", "answer": "[{\"word\":\"montalais\",\"start\":3.07,\"end\":3.7}]"} {"utterance_id": "1221-135766-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0005.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.09, 2.53, "child's"]], "response_mode": "offline", "answer": "[{\"word\":\"child's\",\"start\":2.09,\"end\":2.53}]"} {"utterance_id": "1320-122617-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0032.flac", "task_type": "lexical_detection", "query": "responsible", "prompt": "When the utterance is complete, report all instances of the word \"responsible\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.21, 9.96, "responsible"]], "response_mode": "offline", "answer": "[{\"word\":\"responsible\",\"start\":9.21,\"end\":9.96}]"} {"utterance_id": "4992-41806-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.93, 1.47, "bill harmon"]], "response_mode": "offline", "answer": "[{\"word\":\"bill harmon\",\"start\":0.93,\"end\":1.47}]"} {"utterance_id": "5683-32879-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0025.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.91, 1.33, "dorcas"]], "response_mode": "offline", "answer": "[{\"word\":\"dorcas\",\"start\":0.91,\"end\":1.33}]"} {"utterance_id": "3575-170457-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0036.flac", "task_type": "lexical_counting", "query": "true", "prompt": "Listen to the entire clip, then report the count sequence for the word \"true\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.22, 10.48, "one"], [24.01, 24.47, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":10.22,\"end\":10.48},{\"word\":\"two\",\"start\":24.01,\"end\":24.47}]"} {"utterance_id": "6930-75918-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0009.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.84, 4.21, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":3.84,\"end\":4.21}]"} {"utterance_id": "4507-16021-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0013.flac", "task_type": "lexical_detection", "query": "horrible", "prompt": "When the utterance is complete, report all instances of the word \"horrible\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.73, 8.49, "horrible"]], "response_mode": "offline", "answer": "[{\"word\":\"horrible\",\"start\":7.73,\"end\":8.49}]"} {"utterance_id": "3729-6852-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0002.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.6, 6.84, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":6.6,\"end\":6.84}]"} {"utterance_id": "5142-36377-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0005.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.52, 3.02, "brothers"]], "response_mode": "offline", "answer": "[{\"word\":\"brothers\",\"start\":2.52,\"end\":3.02}]"} {"utterance_id": "8230-279154-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0013.flac", "task_type": "lexical_detection", "query": "belief", "prompt": "After the audio ends, extract all occurrences of the word \"belief\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.77, 9.2, "belief"]], "response_mode": "offline", "answer": "[{\"word\":\"belief\",\"start\":8.77,\"end\":9.2}]"} {"utterance_id": "1188-133604-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0020.flac", "task_type": "lexical_detection", "query": "create", "prompt": "When the utterance is complete, report all instances of the word \"create\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.86, 6.37, "create"]], "response_mode": "offline", "answer": "[{\"word\":\"create\",\"start\":5.86,\"end\":6.37}]"} {"utterance_id": "237-134500-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0014.flac", "task_type": "lexical_detection", "query": "toward", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"toward\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.53, 2.81, "toward"]], "response_mode": "offline", "answer": "[{\"word\":\"toward\",\"start\":2.53,\"end\":2.81}]"} {"utterance_id": "7021-79730-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0004.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Once the audio has finished, identify every instance of a family relation. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.49, 0.93, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":0.49,\"end\":0.93}]"} {"utterance_id": "2830-3980-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0009.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.99, 1.57, "apostle"]], "response_mode": "offline", "answer": "[{\"word\":\"apostle\",\"start\":0.99,\"end\":1.57}]"} {"utterance_id": "237-134500-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0027.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.94, 1.29, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":0.94,\"end\":1.29}]"} {"utterance_id": "6829-68771-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0000.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Do not answer until the audio ends. Then list each instance of a money or currency term as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[10.54, 11.04, "dollars"]], "response_mode": "offline", "answer": "[{\"word\":\"dollars\",\"start\":10.54,\"end\":11.04}]"} {"utterance_id": "1188-133604-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0026.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.91, 3.5, "wilson"], [18.2, 18.74, "apollo"]], "response_mode": "offline", "answer": "[{\"word\":\"wilson\",\"start\":2.91,\"end\":3.5},{\"word\":\"apollo\",\"start\":18.2,\"end\":18.74}]"} {"utterance_id": "1320-122612-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0011.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.75, 8.19, "scout"]], "response_mode": "offline", "answer": "[{\"word\":\"scout\",\"start\":7.75,\"end\":8.19}]"} {"utterance_id": "5142-33396-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0010.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a body part: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[1.32, 1.68, "one"], [2.81, 3.23, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.32,\"end\":1.68},{\"word\":\"two\",\"start\":2.81,\"end\":3.23}]"} {"utterance_id": "7176-88083-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0016.flac", "task_type": "lexical_detection", "query": "glided", "prompt": "When the utterance is complete, report all instances of the word \"glided\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.55, 1.94, "glided"]], "response_mode": "offline", "answer": "[{\"word\":\"glided\",\"start\":1.55,\"end\":1.94}]"} {"utterance_id": "8463-294828-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0021.flac", "task_type": "lexical_detection", "query": "slightly", "prompt": "After listening to the whole recording, output every instance of the word \"slightly\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.77, 1.21, "slightly"]], "response_mode": "offline", "answer": "[{\"word\":\"slightly\",\"start\":0.77,\"end\":1.21}]"} {"utterance_id": "6829-68769-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0051.flac", "task_type": "lexical_detection", "query": "grim", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"grim\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.61, 0.89, "grim"]], "response_mode": "offline", "answer": "[{\"word\":\"grim\",\"start\":0.61,\"end\":0.89}]"} {"utterance_id": "7176-92135-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0017.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.97, 4.14, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.97,\"end\":4.14}]"} {"utterance_id": "1284-1181-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0013.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.17, 0.56, "ojo"]], "response_mode": "offline", "answer": "[{\"word\":\"ojo\",\"start\":0.17,\"end\":0.56}]"} {"utterance_id": "1188-133604-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0032.flac", "task_type": "lexical_detection", "query": "pencil", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"pencil\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.34, 1.72, "pencil"]], "response_mode": "offline", "answer": "[{\"word\":\"pencil\",\"start\":1.34,\"end\":1.72}]"} {"utterance_id": "4077-13754-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0012.flac", "task_type": "semantic_counting", "query": "organization", "prompt": "Listen to the entire clip, then report the count sequence for an organization name as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.15, 4.55, "one"], [15.73, 16.08, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.15,\"end\":4.55},{\"word\":\"two\",\"start\":15.73,\"end\":16.08}]"} {"utterance_id": "1320-122612-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0011.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.36, 7.54, "eye"]], "response_mode": "offline", "answer": "[{\"word\":\"eye\",\"start\":7.36,\"end\":7.54}]"} {"utterance_id": "8463-287645-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0010.flac", "task_type": "lexical_detection", "query": "beating", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"beating\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.64, 3.01, "beating"]], "response_mode": "offline", "answer": "[{\"word\":\"beating\",\"start\":2.64,\"end\":3.01}]"} {"utterance_id": "1089-134686-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0028.flac", "task_type": "semantic_detection", "query": "date", "prompt": "You will hear a complete utterance. When it ends, list every instance of a date you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.89, 2.47, "wednesday"], [6.72, 7.43, "saturday"]], "response_mode": "offline", "answer": "[{\"word\":\"wednesday\",\"start\":1.89,\"end\":2.47},{\"word\":\"saturday\",\"start\":6.72,\"end\":7.43}]"} {"utterance_id": "908-31957-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0019.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "Once the audio has finished, identify every instance of a disease name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.78, 3.42, "sickness"]], "response_mode": "offline", "answer": "[{\"word\":\"sickness\",\"start\":2.78,\"end\":3.42}]"} {"utterance_id": "6930-76324-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0015.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[8.11, 8.64, "cynthia"]], "response_mode": "offline", "answer": "[{\"word\":\"cynthia\",\"start\":8.11,\"end\":8.64}]"} {"utterance_id": "8463-294825-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0018.flac", "task_type": "lexical_detection", "query": "three", "prompt": "After the audio ends, extract all occurrences of the word \"three\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.92, 2.17, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":1.92,\"end\":2.17}]"} {"utterance_id": "8463-294828-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0012.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.04, 4.66, "forty"]], "response_mode": "offline", "answer": "[{\"word\":\"forty\",\"start\":4.04,\"end\":4.66}]"} {"utterance_id": "4970-29095-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0038.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.84, "ruth"], [2.14, 2.45, "philip"]], "response_mode": "offline", "answer": "[{\"word\":\"ruth\",\"start\":0.53,\"end\":0.84},{\"word\":\"philip\",\"start\":2.14,\"end\":2.45}]"} {"utterance_id": "7021-85628-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0003.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.63, 9.13, "anders"]], "response_mode": "offline", "answer": "[{\"word\":\"anders\",\"start\":8.63,\"end\":9.13}]"} {"utterance_id": "4970-29095-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0024.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a family relation with fields word, start, end (seconds). If none, reply [].", "labels": [[1.26, 1.6, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":1.26,\"end\":1.6}]"} {"utterance_id": "4446-2275-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0029.flac", "task_type": "lexical_detection", "query": "please", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"please\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.34, 0.63, "please"]], "response_mode": "offline", "answer": "[{\"word\":\"please\",\"start\":0.34,\"end\":0.63}]"} {"utterance_id": "1995-1837-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0019.flac", "task_type": "lexical_detection", "query": "zora", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"zora\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.56, 5.23, "zora"]], "response_mode": "offline", "answer": "[{\"word\":\"zora\",\"start\":4.56,\"end\":5.23}]"} {"utterance_id": "672-122797-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0054.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Once the audio has finished, identify every instance of a plant name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.38, 3.86, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":3.38,\"end\":3.86}]"} {"utterance_id": "4077-13751-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0013.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.47, 4.07, "scribe"]], "response_mode": "offline", "answer": "[{\"word\":\"scribe\",\"start\":3.47,\"end\":4.07}]"} {"utterance_id": "3570-5694-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0015.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.5, 3.05, "retainer"]], "response_mode": "offline", "answer": "[{\"word\":\"retainer\",\"start\":2.5,\"end\":3.05}]"} {"utterance_id": "4077-13754-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0012.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Once the audio has finished, list each occurrence of a number as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[6.38, 6.71, "one"], [8.65, 9.22, "two"], [20.53, 21.02, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.38,\"end\":6.71},{\"word\":\"two\",\"start\":8.65,\"end\":9.22},{\"word\":\"three\",\"start\":20.53,\"end\":21.02}]"} {"utterance_id": "237-134500-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0028.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.84, 1.49, "alexandra"]], "response_mode": "offline", "answer": "[{\"word\":\"alexandra\",\"start\":0.84,\"end\":1.49}]"} {"utterance_id": "4446-2273-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0009.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a family relation with fields word, start, end (seconds). If none, reply [].", "labels": [[3.32, 3.87, "mother's"]], "response_mode": "offline", "answer": "[{\"word\":\"mother's\",\"start\":3.32,\"end\":3.87}]"} {"utterance_id": "121-121726-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0003.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a disease name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.51, 1.67, "hay fever"]], "response_mode": "offline", "answer": "[{\"word\":\"hay fever\",\"start\":0.51,\"end\":1.67}]"} {"utterance_id": "3570-5696-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0002.flac", "task_type": "lexical_detection", "query": "favor", "prompt": "Process the full audio offline. Find every instance of the word \"favor\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.55, 3.91, "favor"]], "response_mode": "offline", "answer": "[{\"word\":\"favor\",\"start\":3.55,\"end\":3.91}]"} {"utterance_id": "5683-32879-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0024.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.62, 1.11, "rachel"]], "response_mode": "offline", "answer": "[{\"word\":\"rachel\",\"start\":0.62,\"end\":1.11}]"} {"utterance_id": "1221-135767-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0023.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[15.47, 15.81, "bull"]], "response_mode": "offline", "answer": "[{\"word\":\"bull\",\"start\":15.47,\"end\":15.81}]"} {"utterance_id": "1089-134691-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0010.flac", "task_type": "lexical_detection", "query": "brother", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"brother\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.4, 0.64, "brother"], [2.08, 2.35, "brother"]], "response_mode": "offline", "answer": "[{\"word\":\"brother\",\"start\":0.4,\"end\":0.64},{\"word\":\"brother\",\"start\":2.08,\"end\":2.35}]"} {"utterance_id": "1995-1836-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0003.flac", "task_type": "lexical_counting", "query": "intelligence", "prompt": "When the utterance is complete, report how the word \"intelligence\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.35, 5.06, "one"], [6.46, 7.11, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.35,\"end\":5.06},{\"word\":\"two\",\"start\":6.46,\"end\":7.11}]"} {"utterance_id": "5683-32879-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0020.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.77, 4.23, "rachel"]], "response_mode": "offline", "answer": "[{\"word\":\"rachel\",\"start\":3.77,\"end\":4.23}]"} {"utterance_id": "8555-292519-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0007.flac", "task_type": "lexical_detection", "query": "sky", "prompt": "When the utterance is complete, report all instances of the word \"sky\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.98, 2.78, "sky"]], "response_mode": "offline", "answer": "[{\"word\":\"sky\",\"start\":1.98,\"end\":2.78}]"} {"utterance_id": "6829-68771-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0029.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the full audio. After it finishes, find every instance of a family relation. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.95, 3.38, "cousins"]], "response_mode": "offline", "answer": "[{\"word\":\"cousins\",\"start\":2.95,\"end\":3.38}]"} {"utterance_id": "6930-81414-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0020.flac", "task_type": "lexical_detection", "query": "means", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"means\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.36, 2.89, "means"]], "response_mode": "offline", "answer": "[{\"word\":\"means\",\"start\":2.36,\"end\":2.89}]"} {"utterance_id": "2830-3980-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0006.flac", "task_type": "lexical_detection", "query": "boasting", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"boasting\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.1, 1.66, "boasting"]], "response_mode": "offline", "answer": "[{\"word\":\"boasting\",\"start\":1.1,\"end\":1.66}]"} {"utterance_id": "2094-142345-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0043.flac", "task_type": "lexical_detection", "query": "meant", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"meant\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.99, 6.2, "meant"]], "response_mode": "offline", "answer": "[{\"word\":\"meant\",\"start\":5.99,\"end\":6.2}]"} {"utterance_id": "908-31957-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0020.flac", "task_type": "lexical_detection", "query": "thanks", "prompt": "Once the audio has finished, identify every instance of the word \"thanks\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.67, 4.16, "thanks"]], "response_mode": "offline", "answer": "[{\"word\":\"thanks\",\"start\":3.67,\"end\":4.16}]"} {"utterance_id": "7021-79740-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0005.flac", "task_type": "lexical_detection", "query": "glad", "prompt": "When the utterance is complete, report all instances of the word \"glad\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.37, 2.03, "glad"]], "response_mode": "offline", "answer": "[{\"word\":\"glad\",\"start\":1.37,\"end\":2.03}]"} {"utterance_id": "1284-1180-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0002.flac", "task_type": "lexical_detection", "query": "boots", "prompt": "Process the full audio offline. Find every instance of the word \"boots\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.52, 2.97, "boots"]], "response_mode": "offline", "answer": "[{\"word\":\"boots\",\"start\":2.52,\"end\":2.97}]"} {"utterance_id": "2961-961-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0022.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the full audio. After it finishes, find every instance of an ordinal number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.74, 5.32, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":4.74,\"end\":5.32}]"} {"utterance_id": "4507-16021-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0005.flac", "task_type": "lexical_detection", "query": "understood", "prompt": "After listening to the whole recording, output every instance of the word \"understood\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.15, 1.88, "understood"]], "response_mode": "offline", "answer": "[{\"word\":\"understood\",\"start\":1.15,\"end\":1.88}]"} {"utterance_id": "4507-16021-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0017.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.58, 2.52, "philologist"], [6.02, 6.76, "philosopher"]], "response_mode": "offline", "answer": "[{\"word\":\"philologist\",\"start\":1.58,\"end\":2.52},{\"word\":\"philosopher\",\"start\":6.02,\"end\":6.76}]"} {"utterance_id": "7176-92135-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0014.flac", "task_type": "semantic_counting", "query": "ordinal", "prompt": "Listen to the audio in full, then emit a JSON array of count events for an ordinal number (word, start, end). If none, reply [].", "labels": [[1.1, 1.61, "one"], [5.41, 6.07, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.1,\"end\":1.61},{\"word\":\"two\",\"start\":5.41,\"end\":6.07}]"} {"utterance_id": "6829-68771-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0012.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[7.25, 7.65, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.25,\"end\":7.65}]"} {"utterance_id": "1284-134647-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0005.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.36, 7.84, "bishops"]], "response_mode": "offline", "answer": "[{\"word\":\"bishops\",\"start\":7.36,\"end\":7.84}]"} {"utterance_id": "1089-134686-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0004.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.49, 5.07, "husband"]], "response_mode": "offline", "answer": "[{\"word\":\"husband\",\"start\":4.49,\"end\":5.07}]"} {"utterance_id": "1089-134691-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0007.flac", "task_type": "lexical_detection", "query": "bridge", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"bridge\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.24, 1.61, "bridge"]], "response_mode": "offline", "answer": "[{\"word\":\"bridge\",\"start\":1.24,\"end\":1.61}]"} {"utterance_id": "7127-75947-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0010.flac", "task_type": "lexical_detection", "query": "perceived", "prompt": "Process the full audio offline. Find every instance of the word \"perceived\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.86, 1.52, "perceived"]], "response_mode": "offline", "answer": "[{\"word\":\"perceived\",\"start\":0.86,\"end\":1.52}]"} {"utterance_id": "7021-85628-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0018.flac", "task_type": "lexical_detection", "query": "everybody", "prompt": "After the audio ends, extract all occurrences of the word \"everybody\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.8, 6.57, "everybody"]], "response_mode": "offline", "answer": "[{\"word\":\"everybody\",\"start\":5.8,\"end\":6.57}]"} {"utterance_id": "4077-13751-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.77, 1.05, "mormon"]], "response_mode": "offline", "answer": "[{\"word\":\"mormon\",\"start\":0.77,\"end\":1.05}]"} {"utterance_id": "3570-5696-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0004.flac", "task_type": "lexical_detection", "query": "salient", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"salient\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.42, 0.83, "salient"]], "response_mode": "offline", "answer": "[{\"word\":\"salient\",\"start\":0.42,\"end\":0.83}]"} {"utterance_id": "7729-102255-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0015.flac", "task_type": "semantic_counting", "query": "organization", "prompt": "Listen to the entire clip, then report the count sequence for an organization name as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.9, 1.36, "one"], [8.79, 9.18, "two"], [12.46, 13.48, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.9,\"end\":1.36},{\"word\":\"two\",\"start\":8.79,\"end\":9.18},{\"word\":\"three\",\"start\":12.46,\"end\":13.48}]"} {"utterance_id": "4446-2271-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0005.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.94, 1.21, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":0.94,\"end\":1.21}]"} {"utterance_id": "2300-131720-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0014.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.34, 1.2, "mister edison"]], "response_mode": "offline", "answer": "[{\"word\":\"mister edison\",\"start\":0.34,\"end\":1.2}]"} {"utterance_id": "1188-133604-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0008.flac", "task_type": "lexical_counting", "query": "art", "prompt": "Listen to the entire clip, then report the count sequence for the word \"art\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.19, 3.45, "one"], [11.75, 11.95, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.19,\"end\":3.45},{\"word\":\"two\",\"start\":11.75,\"end\":11.95}]"} {"utterance_id": "672-122797-0055", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0055.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.91, 4.35, "mice"]], "response_mode": "offline", "answer": "[{\"word\":\"mice\",\"start\":3.91,\"end\":4.35}]"} {"utterance_id": "4077-13751-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0001.flac", "task_type": "lexical_detection", "query": "insignificant", "prompt": "After listening to the whole recording, output every instance of the word \"insignificant\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.28, 4.07, "insignificant"]], "response_mode": "offline", "answer": "[{\"word\":\"insignificant\",\"start\":3.28,\"end\":4.07}]"} {"utterance_id": "2094-142345-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0048.flac", "task_type": "lexical_detection", "query": "seating", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"seating\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.73, 2.04, "seating"]], "response_mode": "offline", "answer": "[{\"word\":\"seating\",\"start\":1.73,\"end\":2.04}]"} {"utterance_id": "260-123286-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0028.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Listen to the full audio. Count every occurrence of a body part. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.74, 1.09, "one"], [6.46, 7.0, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.74,\"end\":1.09},{\"word\":\"two\",\"start\":6.46,\"end\":7.0}]"} {"utterance_id": "7021-85628-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0021.flac", "task_type": "lexical_detection", "query": "rushed", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"rushed\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.85, 3.2, "rushed"]], "response_mode": "offline", "answer": "[{\"word\":\"rushed\",\"start\":2.85,\"end\":3.2}]"} {"utterance_id": "1580-141083-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0029.flac", "task_type": "lexical_detection", "query": "midst", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"midst\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.58, 0.87, "midst"]], "response_mode": "offline", "answer": "[{\"word\":\"midst\",\"start\":0.58,\"end\":0.87}]"} {"utterance_id": "8555-284449-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0000.flac", "task_type": "lexical_detection", "query": "clashing", "prompt": "Process the full audio offline. Find every instance of the word \"clashing\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.65, 7.15, "clashing"]], "response_mode": "offline", "answer": "[{\"word\":\"clashing\",\"start\":6.65,\"end\":7.15}]"} {"utterance_id": "8224-274381-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0006.flac", "task_type": "lexical_detection", "query": "montrose", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"montrose\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.68, 15.46, "montrose"]], "response_mode": "offline", "answer": "[{\"word\":\"montrose\",\"start\":14.68,\"end\":15.46}]"} {"utterance_id": "4992-41797-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0009.flac", "task_type": "lexical_detection", "query": "henry", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"henry\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.43, 0.82, "henry"]], "response_mode": "offline", "answer": "[{\"word\":\"henry\",\"start\":0.43,\"end\":0.82}]"} {"utterance_id": "8455-210777-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0051.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.79, 6.3, "captain"]], "response_mode": "offline", "answer": "[{\"word\":\"captain\",\"start\":5.79,\"end\":6.3}]"} {"utterance_id": "4077-13751-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0003.flac", "task_type": "semantic_detection", "query": "city", "prompt": "After the audio ends, extract all occurrences of a city name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.34, 1.93, "hamlet"]], "response_mode": "offline", "answer": "[{\"word\":\"hamlet\",\"start\":1.34,\"end\":1.93}]"} {"utterance_id": "1320-122617-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0016.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen carefully to the entire clip, then report every instance of a family relation as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.13, 3.58, "brothers"]], "response_mode": "offline", "answer": "[{\"word\":\"brothers\",\"start\":3.13,\"end\":3.58}]"} {"utterance_id": "2830-3980-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0002.flac", "task_type": "lexical_detection", "query": "turn", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"turn\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.08, 1.3, "turn"]], "response_mode": "offline", "answer": "[{\"word\":\"turn\",\"start\":1.08,\"end\":1.3}]"} {"utterance_id": "8555-284449-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0000.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.08, 3.55, "queen"]], "response_mode": "offline", "answer": "[{\"word\":\"queen\",\"start\":3.08,\"end\":3.55}]"} {"utterance_id": "8455-210777-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0053.flac", "task_type": "lexical_detection", "query": "expressing", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"expressing\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.38, 2.84, "expressing"]], "response_mode": "offline", "answer": "[{\"word\":\"expressing\",\"start\":2.38,\"end\":2.84}]"} {"utterance_id": "2830-3979-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0005.flac", "task_type": "lexical_detection", "query": "work", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"work\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 0.72, "work"]], "response_mode": "offline", "answer": "[{\"word\":\"work\",\"start\":0.51,\"end\":0.72}]"} {"utterance_id": "1089-134686-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0011.flac", "task_type": "lexical_detection", "query": "recite", "prompt": "After listening to the whole recording, output every instance of the word \"recite\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.84, 4.28, "recite"]], "response_mode": "offline", "answer": "[{\"word\":\"recite\",\"start\":3.84,\"end\":4.28}]"} {"utterance_id": "121-127105-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0030.flac", "task_type": "lexical_detection", "query": "don't", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"don't\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.62, 1.15, "don't"]], "response_mode": "offline", "answer": "[{\"word\":\"don't\",\"start\":0.62,\"end\":1.15}]"} {"utterance_id": "6829-68769-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0022.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.25, 2.6, "kenneth"]], "response_mode": "offline", "answer": "[{\"word\":\"kenneth\",\"start\":2.25,\"end\":2.6}]"} {"utterance_id": "4446-2271-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.27, 1.74, "winifred"]], "response_mode": "offline", "answer": "[{\"word\":\"winifred\",\"start\":1.27,\"end\":1.74}]"} {"utterance_id": "121-121726-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0001.flac", "task_type": "lexical_detection", "query": "tiresome", "prompt": "After the audio ends, extract all occurrences of the word \"tiresome\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.0, 3.75, "tiresome"]], "response_mode": "offline", "answer": "[{\"word\":\"tiresome\",\"start\":3.0,\"end\":3.75}]"} {"utterance_id": "237-126133-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0023.flac", "task_type": "lexical_detection", "query": "owned", "prompt": "After the audio ends, extract all occurrences of the word \"owned\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.44, 2.72, "owned"]], "response_mode": "offline", "answer": "[{\"word\":\"owned\",\"start\":2.44,\"end\":2.72}]"} {"utterance_id": "6829-68769-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0036.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.89, 2.31, "jailer"]], "response_mode": "offline", "answer": "[{\"word\":\"jailer\",\"start\":1.89,\"end\":2.31}]"} {"utterance_id": "3570-5694-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0003.flac", "task_type": "lexical_detection", "query": "custom", "prompt": "Once the audio has finished, identify every instance of the word \"custom\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.93, 4.5, "custom"]], "response_mode": "offline", "answer": "[{\"word\":\"custom\",\"start\":3.93,\"end\":4.5}]"} {"utterance_id": "7729-102255-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0037.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[7.58, 8.14, "families"]], "response_mode": "offline", "answer": "[{\"word\":\"families\",\"start\":7.58,\"end\":8.14}]"} {"utterance_id": "908-157963-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0027.flac", "task_type": "lexical_detection", "query": "cold", "prompt": "After the audio ends, extract all occurrences of the word \"cold\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.61, 2.22, "cold"]], "response_mode": "offline", "answer": "[{\"word\":\"cold\",\"start\":1.61,\"end\":2.22}]"} {"utterance_id": "1284-1181-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0009.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.27, 1.76, "husband's"]], "response_mode": "offline", "answer": "[{\"word\":\"husband's\",\"start\":1.27,\"end\":1.76}]"} {"utterance_id": "121-127105-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0025.flac", "task_type": "lexical_counting", "query": "stay", "prompt": "After the audio ends, tally every instance of the word \"stay\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[2.34, 2.69, "one"], [4.7, 5.33, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.34,\"end\":2.69},{\"word\":\"two\",\"start\":4.7,\"end\":5.33}]"} {"utterance_id": "4446-2273-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0008.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.45, 3.63, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":3.45,\"end\":3.63}]"} {"utterance_id": "4446-2271-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0008.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[1.01, 1.16, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.01,\"end\":1.16}]"} {"utterance_id": "61-70968-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0050.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.06, 3.38, "robin"]], "response_mode": "offline", "answer": "[{\"word\":\"robin\",\"start\":3.06,\"end\":3.38}]"} {"utterance_id": "2830-3980-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0002.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Do not answer until the audio ends. Then list each instance of an ordinal number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.62, 0.98, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":0.62,\"end\":0.98}]"} {"utterance_id": "5142-33396-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0020.flac", "task_type": "lexical_detection", "query": "smells", "prompt": "Once the audio has finished, identify every instance of the word \"smells\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.59, 0.89, "smells"], [2.04, 2.38, "smells"]], "response_mode": "offline", "answer": "[{\"word\":\"smells\",\"start\":0.59,\"end\":0.89},{\"word\":\"smells\",\"start\":2.04,\"end\":2.38}]"} {"utterance_id": "2830-3980-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0019.flac", "task_type": "lexical_detection", "query": "certainty", "prompt": "When the utterance is complete, report all instances of the word \"certainty\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.83, 4.4, "certainty"]], "response_mode": "offline", "answer": "[{\"word\":\"certainty\",\"start\":3.83,\"end\":4.4}]"} {"utterance_id": "5105-28241-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0005.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Listen carefully to the entire clip, then report every instance of a country name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.1, 5.81, "algeria"]], "response_mode": "offline", "answer": "[{\"word\":\"algeria\",\"start\":5.1,\"end\":5.81}]"} {"utterance_id": "1320-122612-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0012.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "After the audio ends, extract all occurrences of an animal name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.22, 4.51, "deer"], [9.38, 10.05, "horses"]], "response_mode": "offline", "answer": "[{\"word\":\"deer\",\"start\":4.22,\"end\":4.51},{\"word\":\"horses\",\"start\":9.38,\"end\":10.05}]"} {"utterance_id": "6829-68771-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0009.flac", "task_type": "lexical_detection", "query": "entertainment", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"entertainment\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.22, 6.0, "entertainment"]], "response_mode": "offline", "answer": "[{\"word\":\"entertainment\",\"start\":5.22,\"end\":6.0}]"} {"utterance_id": "2830-3980-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0059.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 0.78, "paul"]], "response_mode": "offline", "answer": "[{\"word\":\"paul\",\"start\":0.52,\"end\":0.78}]"} {"utterance_id": "2830-3980-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0004.flac", "task_type": "lexical_detection", "query": "church", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"church\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.67, 1.96, "church"]], "response_mode": "offline", "answer": "[{\"word\":\"church\",\"start\":1.67,\"end\":1.96}]"} {"utterance_id": "5105-28241-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0000.flac", "task_type": "lexical_detection", "query": "excellent", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"excellent\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.73, 2.34, "excellent"]], "response_mode": "offline", "answer": "[{\"word\":\"excellent\",\"start\":1.73,\"end\":2.34}]"} {"utterance_id": "8230-279154-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0022.flac", "task_type": "lexical_detection", "query": "points", "prompt": "When the utterance is complete, report all instances of the word \"points\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.83, 1.21, "points"]], "response_mode": "offline", "answer": "[{\"word\":\"points\",\"start\":0.83,\"end\":1.21}]"} {"utterance_id": "4992-41806-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0014.flac", "task_type": "semantic_counting", "query": "clock_time", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a time of day: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[10.46, 10.76, "one"], [13.41, 13.91, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":10.46,\"end\":10.76},{\"word\":\"two\",\"start\":13.41,\"end\":13.91}]"} {"utterance_id": "8224-274381-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0004.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a number: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.34, 1.08, "one"], [15.25, 16.16, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.34,\"end\":1.08},{\"word\":\"two\",\"start\":15.25,\"end\":16.16}]"} {"utterance_id": "8463-287645-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0011.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Process the full audio offline. Find every instance of a color name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.17, 3.46, "white"]], "response_mode": "offline", "answer": "[{\"word\":\"white\",\"start\":3.17,\"end\":3.46}]"} {"utterance_id": "908-157963-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0003.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.45, 2.08, "children"]], "response_mode": "offline", "answer": "[{\"word\":\"children\",\"start\":1.45,\"end\":2.08}]"} {"utterance_id": "2830-3980-0066", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0066.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.58, 0.97, "christ"]], "response_mode": "offline", "answer": "[{\"word\":\"christ\",\"start\":0.58,\"end\":0.97}]"} {"utterance_id": "5142-33396-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0015.flac", "task_type": "lexical_detection", "query": "boat", "prompt": "Process the full audio offline. Find every instance of the word \"boat\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.41, 0.69, "boat"]], "response_mode": "offline", "answer": "[{\"word\":\"boat\",\"start\":0.41,\"end\":0.69}]"} {"utterance_id": "8230-279154-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0016.flac", "task_type": "lexical_detection", "query": "actual", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"actual\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.71, 1.16, "actual"]], "response_mode": "offline", "answer": "[{\"word\":\"actual\",\"start\":0.71,\"end\":1.16}]"} {"utterance_id": "237-134493-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0014.flac", "task_type": "lexical_detection", "query": "you've", "prompt": "After listening to the whole recording, output every instance of the word \"you've\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.56, 1.74, "you've"]], "response_mode": "offline", "answer": "[{\"word\":\"you've\",\"start\":1.56,\"end\":1.74}]"} {"utterance_id": "7176-92135-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0031.flac", "task_type": "lexical_detection", "query": "sentences", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"sentences\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.0, 4.63, "sentences"]], "response_mode": "offline", "answer": "[{\"word\":\"sentences\",\"start\":4.0,\"end\":4.63}]"} {"utterance_id": "1089-134686-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0021.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "You will hear a complete utterance. When it ends, list every instance of a family relation you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.04, 5.5, "child"]], "response_mode": "offline", "answer": "[{\"word\":\"child\",\"start\":5.04,\"end\":5.5}]"} {"utterance_id": "672-122797-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0032.flac", "task_type": "lexical_detection", "query": "ladies", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"ladies\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.79, 1.43, "ladies"]], "response_mode": "offline", "answer": "[{\"word\":\"ladies\",\"start\":0.79,\"end\":1.43}]"} {"utterance_id": "260-123288-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0007.flac", "task_type": "lexical_detection", "query": "increased", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"increased\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.63, 3.15, "increased"]], "response_mode": "offline", "answer": "[{\"word\":\"increased\",\"start\":2.63,\"end\":3.15}]"} {"utterance_id": "5142-33396-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0045.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.66, 1.03, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":0.66,\"end\":1.03}]"} {"utterance_id": "2830-3980-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0027.flac", "task_type": "lexical_detection", "query": "brethren", "prompt": "After the audio ends, extract all occurrences of the word \"brethren\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.74, 1.2, "brethren"]], "response_mode": "offline", "answer": "[{\"word\":\"brethren\",\"start\":0.74,\"end\":1.2}]"} {"utterance_id": "8455-210777-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0040.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.62, 2.67, "sir ferdinando"]], "response_mode": "offline", "answer": "[{\"word\":\"sir ferdinando\",\"start\":1.62,\"end\":2.67}]"} {"utterance_id": "5142-33396-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0039.flac", "task_type": "lexical_detection", "query": "named", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"named\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.39, 0.66, "named"]], "response_mode": "offline", "answer": "[{\"word\":\"named\",\"start\":0.39,\"end\":0.66}]"} {"utterance_id": "260-123440-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0001.flac", "task_type": "lexical_detection", "query": "alice", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"alice\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.92, 1.56, "alice"]], "response_mode": "offline", "answer": "[{\"word\":\"alice\",\"start\":0.92,\"end\":1.56}]"} {"utterance_id": "237-134500-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0027.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 0.94, "marie's"]], "response_mode": "offline", "answer": "[{\"word\":\"marie's\",\"start\":0.51,\"end\":0.94}]"} {"utterance_id": "5142-33396-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0021.flac", "task_type": "lexical_detection", "query": "frolic", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"frolic\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.84, 3.37, "frolic"]], "response_mode": "offline", "answer": "[{\"word\":\"frolic\",\"start\":2.84,\"end\":3.37}]"} {"utterance_id": "672-122797-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0007.flac", "task_type": "lexical_detection", "query": "largest", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"largest\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.73, 5.41, "largest"]], "response_mode": "offline", "answer": "[{\"word\":\"largest\",\"start\":4.73,\"end\":5.41}]"} {"utterance_id": "3575-170457-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0031.flac", "task_type": "lexical_detection", "query": "seven", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"seven\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.42, 2.75, "seven"]], "response_mode": "offline", "answer": "[{\"word\":\"seven\",\"start\":2.42,\"end\":2.75}]"} {"utterance_id": "4446-2273-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0016.flac", "task_type": "lexical_counting", "query": "like", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"like\" with word/start/end. If none, reply [].", "labels": [[1.33, 1.6, "one"], [3.77, 3.95, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.33,\"end\":1.6},{\"word\":\"two\",\"start\":3.77,\"end\":3.95}]"} {"utterance_id": "4446-2273-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0015.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.6, 3.86, "son"]], "response_mode": "offline", "answer": "[{\"word\":\"son\",\"start\":3.6,\"end\":3.86}]"} {"utterance_id": "2300-131720-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0034.flac", "task_type": "lexical_detection", "query": "operation", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"operation\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.73, 2.35, "operation"]], "response_mode": "offline", "answer": "[{\"word\":\"operation\",\"start\":1.73,\"end\":2.35}]"} {"utterance_id": "6930-76324-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0020.flac", "task_type": "lexical_detection", "query": "vast", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"vast\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.57, 4.14, "vast"]], "response_mode": "offline", "answer": "[{\"word\":\"vast\",\"start\":3.57,\"end\":4.14}]"} {"utterance_id": "7021-85628-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0021.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.1, 2.56, "hands"]], "response_mode": "offline", "answer": "[{\"word\":\"hands\",\"start\":2.1,\"end\":2.56}]"} {"utterance_id": "5683-32865-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.02, 2.03, "captain lake"]], "response_mode": "offline", "answer": "[{\"word\":\"captain lake\",\"start\":1.02,\"end\":2.03}]"} {"utterance_id": "8463-287645-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0005.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a time of day with fields word, start, end (seconds). If none, reply [].", "labels": [[5.43, 5.74, "night"]], "response_mode": "offline", "answer": "[{\"word\":\"night\",\"start\":5.43,\"end\":5.74}]"} {"utterance_id": "8230-279154-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0003.flac", "task_type": "lexical_detection", "query": "evidence", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"evidence\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.89, 1.39, "evidence"]], "response_mode": "offline", "answer": "[{\"word\":\"evidence\",\"start\":0.89,\"end\":1.39}]"} {"utterance_id": "4970-29093-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0022.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a family relation with fields word, start, end (seconds). If none, reply [].", "labels": [[1.06, 1.41, "uncle"]], "response_mode": "offline", "answer": "[{\"word\":\"uncle\",\"start\":1.06,\"end\":1.41}]"} {"utterance_id": "4970-29095-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0006.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.68, 1.18, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":0.68,\"end\":1.18}]"} {"utterance_id": "2830-3980-0065", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0065.flac", "task_type": "lexical_detection", "query": "answers", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"answers\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.6, 1.16, "answers"]], "response_mode": "offline", "answer": "[{\"word\":\"answers\",\"start\":0.6,\"end\":1.16}]"} {"utterance_id": "6930-81414-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0004.flac", "task_type": "lexical_detection", "query": "evil", "prompt": "After listening to the whole recording, output every instance of the word \"evil\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.24, 1.69, "evil"]], "response_mode": "offline", "answer": "[{\"word\":\"evil\",\"start\":1.24,\"end\":1.69}]"} {"utterance_id": "7176-92135-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0045.flac", "task_type": "lexical_detection", "query": "upper", "prompt": "Once the audio has finished, identify every instance of the word \"upper\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.84, 4.14, "upper"]], "response_mode": "offline", "answer": "[{\"word\":\"upper\",\"start\":3.84,\"end\":4.14}]"} {"utterance_id": "5142-36377-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0012.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.86, 1.72, "mister jago"]], "response_mode": "offline", "answer": "[{\"word\":\"mister jago\",\"start\":0.86,\"end\":1.72}]"} {"utterance_id": "1580-141083-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0018.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.5, 1.32, "mister soames"]], "response_mode": "offline", "answer": "[{\"word\":\"mister soames\",\"start\":0.5,\"end\":1.32}]"} {"utterance_id": "6930-75918-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0008.flac", "task_type": "semantic_detection", "query": "city", "prompt": "After the audio ends, extract all occurrences of a city name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.47, 2.07, "buckingham"]], "response_mode": "offline", "answer": "[{\"word\":\"buckingham\",\"start\":1.47,\"end\":2.07}]"} {"utterance_id": "1580-141084-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0029.flac", "task_type": "lexical_detection", "query": "glanced", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"glanced\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.53, 2.04, "glanced"]], "response_mode": "offline", "answer": "[{\"word\":\"glanced\",\"start\":1.53,\"end\":2.04}]"} {"utterance_id": "3575-170457-0055", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0055.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.29, 3.63, "anne's"]], "response_mode": "offline", "answer": "[{\"word\":\"anne's\",\"start\":3.29,\"end\":3.63}]"} {"utterance_id": "4077-13754-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0014.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the full audio. After it finishes, find every instance of an organization name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.75, 3.31, "courts"]], "response_mode": "offline", "answer": "[{\"word\":\"courts\",\"start\":2.75,\"end\":3.31}]"} {"utterance_id": "7021-79759-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79759/7021-79759-0004.flac", "task_type": "lexical_detection", "query": "conceptions", "prompt": "Process the full audio offline. Find every instance of the word \"conceptions\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.83, 10.65, "conceptions"]], "response_mode": "offline", "answer": "[{\"word\":\"conceptions\",\"start\":9.83,\"end\":10.65}]"} {"utterance_id": "8555-284447-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0019.flac", "task_type": "lexical_detection", "query": "tumbled", "prompt": "When the utterance is complete, report all instances of the word \"tumbled\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.35, 6.64, "tumbled"]], "response_mode": "offline", "answer": "[{\"word\":\"tumbled\",\"start\":6.35,\"end\":6.64}]"} {"utterance_id": "4446-2273-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0022.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a money or currency term with fields word, start, end (seconds). If none, reply [].", "labels": [[3.15, 3.55, "money"]], "response_mode": "offline", "answer": "[{\"word\":\"money\",\"start\":3.15,\"end\":3.55}]"} {"utterance_id": "1188-133604-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0027.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.41, 3.75, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":3.41,\"end\":3.75}]"} {"utterance_id": "5142-33396-0068", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0068.flac", "task_type": "lexical_detection", "query": "thrall", "prompt": "When the utterance is complete, report all instances of the word \"thrall\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.97, 2.37, "thrall"]], "response_mode": "offline", "answer": "[{\"word\":\"thrall\",\"start\":1.97,\"end\":2.37}]"} {"utterance_id": "1221-135766-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0001.flac", "task_type": "semantic_counting", "query": "kinship", "prompt": "Once the audio has finished, list each occurrence of a family relation as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[5.5, 6.02, "one"], [9.81, 10.31, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.5,\"end\":6.02},{\"word\":\"two\",\"start\":9.81,\"end\":10.31}]"} {"utterance_id": "8463-287645-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0009.flac", "task_type": "lexical_detection", "query": "please", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"please\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.9, 3.28, "please"]], "response_mode": "offline", "answer": "[{\"word\":\"please\",\"start\":2.9,\"end\":3.28}]"} {"utterance_id": "672-122797-0064", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0064.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Do not answer until the audio ends. Then list each instance of a plant name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.35, 4.85, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":4.35,\"end\":4.85}]"} {"utterance_id": "1580-141084-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0026.flac", "task_type": "lexical_detection", "query": "resolve", "prompt": "After the audio ends, extract all occurrences of the word \"resolve\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.28, 4.63, "resolve"]], "response_mode": "offline", "answer": "[{\"word\":\"resolve\",\"start\":4.28,\"end\":4.63}]"} {"utterance_id": "4446-2275-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0005.flac", "task_type": "lexical_detection", "query": "bones", "prompt": "Once the audio has finished, identify every instance of the word \"bones\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.04, 1.49, "bones"]], "response_mode": "offline", "answer": "[{\"word\":\"bones\",\"start\":1.04,\"end\":1.49}]"} {"utterance_id": "260-123440-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0009.flac", "task_type": "lexical_detection", "query": "twenty", "prompt": "Once the audio has finished, identify every instance of the word \"twenty\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.39, 1.99, "twenty"]], "response_mode": "offline", "answer": "[{\"word\":\"twenty\",\"start\":1.39,\"end\":1.99}]"} {"utterance_id": "1089-134691-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0002.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Do not answer until the audio ends. Then list each instance of an animal name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.95, 2.41, "bull"]], "response_mode": "offline", "answer": "[{\"word\":\"bull\",\"start\":1.95,\"end\":2.41}]"} {"utterance_id": "2300-131720-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0020.flac", "task_type": "lexical_detection", "query": "specially", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"specially\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.56, 1.16, "specially"]], "response_mode": "offline", "answer": "[{\"word\":\"specially\",\"start\":0.56,\"end\":1.16}]"} {"utterance_id": "8455-210777-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0050.flac", "task_type": "lexical_detection", "query": "remains", "prompt": "Once the audio has finished, identify every instance of the word \"remains\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.39, 1.76, "remains"]], "response_mode": "offline", "answer": "[{\"word\":\"remains\",\"start\":1.39,\"end\":1.76}]"} {"utterance_id": "3729-6852-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0043.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "You will hear a complete utterance. When it ends, list every instance of a state or province name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.75, 2.48, "marais"]], "response_mode": "offline", "answer": "[{\"word\":\"marais\",\"start\":1.75,\"end\":2.48}]"} {"utterance_id": "672-122797-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0002.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "When the utterance is complete, report all instances of a plant name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.29, 13.1, "strawberries"]], "response_mode": "offline", "answer": "[{\"word\":\"strawberries\",\"start\":12.29,\"end\":13.1}]"} {"utterance_id": "1284-1181-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.98, "ojo"]], "response_mode": "offline", "answer": "[{\"word\":\"ojo\",\"start\":0.53,\"end\":0.98}]"} {"utterance_id": "7021-85628-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.39, 1.8, "anders"]], "response_mode": "offline", "answer": "[{\"word\":\"anders\",\"start\":1.39,\"end\":1.8}]"} {"utterance_id": "3570-5694-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0011.flac", "task_type": "lexical_detection", "query": "sole", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"sole\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.84, 10.21, "sole"]], "response_mode": "offline", "answer": "[{\"word\":\"sole\",\"start\":9.84,\"end\":10.21}]"} {"utterance_id": "8555-292519-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0012.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Listen carefully to the entire clip, then report every instance of a weather term as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.87, 2.41, "rain"]], "response_mode": "offline", "answer": "[{\"word\":\"rain\",\"start\":1.87,\"end\":2.41}]"} {"utterance_id": "1221-135767-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0013.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.17, 3.85, "hester prynne"]], "response_mode": "offline", "answer": "[{\"word\":\"hester prynne\",\"start\":3.17,\"end\":3.85}]"} {"utterance_id": "8224-274384-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0008.flac", "task_type": "lexical_detection", "query": "fallen", "prompt": "Process the full audio offline. Find every instance of the word \"fallen\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.76, 3.28, "fallen"]], "response_mode": "offline", "answer": "[{\"word\":\"fallen\",\"start\":2.76,\"end\":3.28}]"} {"utterance_id": "237-134493-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0004.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.86, 5.17, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.86,\"end\":5.17}]"} {"utterance_id": "2830-3980-0058", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0058.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Once the audio has finished, list each occurrence of a person's name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[0.51, 0.95, "one"], [1.96, 2.42, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.51,\"end\":0.95},{\"word\":\"two\",\"start\":1.96,\"end\":2.42}]"} {"utterance_id": "4992-41806-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0015.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[7.46, 7.67, "head"]], "response_mode": "offline", "answer": "[{\"word\":\"head\",\"start\":7.46,\"end\":7.67}]"} {"utterance_id": "4077-13751-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0012.flac", "task_type": "lexical_counting", "query": "governor", "prompt": "Listen to the entire clip, then report the count sequence for the word \"governor\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.94, 1.48, "one"], [3.86, 4.39, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.94,\"end\":1.48},{\"word\":\"two\",\"start\":3.86,\"end\":4.39}]"} {"utterance_id": "1284-1181-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0014.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen carefully to the entire clip, then report every instance of a color name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.54, 1.84, "gold"]], "response_mode": "offline", "answer": "[{\"word\":\"gold\",\"start\":1.54,\"end\":1.84}]"} {"utterance_id": "4970-29093-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0012.flac", "task_type": "lexical_detection", "query": "declining", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"declining\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.97, 4.52, "declining"]], "response_mode": "offline", "answer": "[{\"word\":\"declining\",\"start\":3.97,\"end\":4.52}]"} {"utterance_id": "7127-75947-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0017.flac", "task_type": "lexical_detection", "query": "already", "prompt": "Process the full audio offline. Find every instance of the word \"already\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.87, 1.29, "already"]], "response_mode": "offline", "answer": "[{\"word\":\"already\",\"start\":0.87,\"end\":1.29}]"} {"utterance_id": "5142-33396-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0008.flac", "task_type": "lexical_detection", "query": "painted", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"painted\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.21, 0.59, "painted"]], "response_mode": "offline", "answer": "[{\"word\":\"painted\",\"start\":0.21,\"end\":0.59}]"} {"utterance_id": "6930-81414-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0017.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Do not answer until the audio ends. Then list each instance of a time of day as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.71, 2.23, "evening"]], "response_mode": "offline", "answer": "[{\"word\":\"evening\",\"start\":1.71,\"end\":2.23}]"} {"utterance_id": "6930-76324-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.06, 2.51, "cynthia"]], "response_mode": "offline", "answer": "[{\"word\":\"cynthia\",\"start\":2.06,\"end\":2.51}]"} {"utterance_id": "2300-131720-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0000.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a city name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.29, 0.73, "paris"]], "response_mode": "offline", "answer": "[{\"word\":\"paris\",\"start\":0.29,\"end\":0.73}]"} {"utterance_id": "4992-41797-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0002.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "You will hear a complete utterance. When it ends, list every instance of a family relation you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.37, 1.08, "grandfather"]], "response_mode": "offline", "answer": "[{\"word\":\"grandfather\",\"start\":0.37,\"end\":1.08}]"} {"utterance_id": "8463-294825-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0016.flac", "task_type": "lexical_detection", "query": "ounce", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"ounce\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.65, 4.13, "ounce"]], "response_mode": "offline", "answer": "[{\"word\":\"ounce\",\"start\":3.65,\"end\":4.13}]"} {"utterance_id": "237-126133-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0015.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.93, 3.48, "jasper"], [8.33, 8.81, "phronsie"]], "response_mode": "offline", "answer": "[{\"word\":\"jasper\",\"start\":2.93,\"end\":3.48},{\"word\":\"phronsie\",\"start\":8.33,\"end\":8.81}]"} {"utterance_id": "61-70968-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0003.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a family relation with fields word, start, end (seconds). If none, reply [].", "labels": [[1.15, 1.66, "father"], [3.47, 3.98, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":1.15,\"end\":1.66},{\"word\":\"father\",\"start\":3.47,\"end\":3.98}]"} {"utterance_id": "3729-6852-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0033.flac", "task_type": "lexical_detection", "query": "selling", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"selling\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.81, 5.2, "selling"]], "response_mode": "offline", "answer": "[{\"word\":\"selling\",\"start\":4.81,\"end\":5.2}]"} {"utterance_id": "7021-79730-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0008.flac", "task_type": "lexical_detection", "query": "managed", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"managed\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[6.74, 7.47, "managed"]], "response_mode": "offline", "answer": "[{\"word\":\"managed\",\"start\":6.74,\"end\":7.47}]"} {"utterance_id": "1089-134686-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0007.flac", "task_type": "lexical_detection", "query": "reigned", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"reigned\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.62, 3.14, "reigned"]], "response_mode": "offline", "answer": "[{\"word\":\"reigned\",\"start\":2.62,\"end\":3.14}]"} {"utterance_id": "7176-92135-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0028.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "When the utterance is complete, report all instances of a time of day including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.62, 1.94, "morning"]], "response_mode": "offline", "answer": "[{\"word\":\"morning\",\"start\":1.62,\"end\":1.94}]"} {"utterance_id": "237-126133-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.54, 6.89, "turner"], [8.04, 8.45, "polly's"]], "response_mode": "offline", "answer": "[{\"word\":\"turner\",\"start\":6.54,\"end\":6.89},{\"word\":\"polly's\",\"start\":8.04,\"end\":8.45}]"} {"utterance_id": "2961-960-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0016.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.71, 1.2, "plato"]], "response_mode": "offline", "answer": "[{\"word\":\"plato\",\"start\":0.71,\"end\":1.2}]"} {"utterance_id": "1284-1181-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0011.flac", "task_type": "lexical_detection", "query": "placing", "prompt": "Once the audio has finished, identify every instance of the word \"placing\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.19, 3.66, "placing"]], "response_mode": "offline", "answer": "[{\"word\":\"placing\",\"start\":3.19,\"end\":3.66}]"} {"utterance_id": "3575-170457-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0005.flac", "task_type": "lexical_detection", "query": "clergyman's", "prompt": "After listening to the whole recording, output every instance of the word \"clergyman's\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.33, 1.92, "clergyman's"]], "response_mode": "offline", "answer": "[{\"word\":\"clergyman's\",\"start\":1.33,\"end\":1.92}]"} {"utterance_id": "7729-102255-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0028.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "When the utterance is complete, report all instances of an organization name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.13, 3.43, "free state hotel"]], "response_mode": "offline", "answer": "[{\"word\":\"free state hotel\",\"start\":2.13,\"end\":3.43}]"} {"utterance_id": "1089-134691-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0002.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After listening to the whole recording, output every instance of an organization name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.41, 9.78, "police"]], "response_mode": "offline", "answer": "[{\"word\":\"police\",\"start\":9.41,\"end\":9.78}]"} {"utterance_id": "8455-210777-0064", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0064.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen carefully to the entire clip, then report every instance of an organization name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.06, 3.48, "government"]], "response_mode": "offline", "answer": "[{\"word\":\"government\",\"start\":3.06,\"end\":3.48}]"} {"utterance_id": "2300-131720-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0013.flac", "task_type": "lexical_detection", "query": "realized", "prompt": "After the audio ends, extract all occurrences of the word \"realized\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.51, 10.24, "realized"]], "response_mode": "offline", "answer": "[{\"word\":\"realized\",\"start\":9.51,\"end\":10.24}]"} {"utterance_id": "3729-6852-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0006.flac", "task_type": "lexical_detection", "query": "compelled", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"compelled\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[7.37, 7.88, "compelled"]], "response_mode": "offline", "answer": "[{\"word\":\"compelled\",\"start\":7.37,\"end\":7.88}]"} {"utterance_id": "1320-122612-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0010.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a color name with fields word, start, end (seconds). If none, reply [].", "labels": [[7.33, 7.7, "dark"]], "response_mode": "offline", "answer": "[{\"word\":\"dark\",\"start\":7.33,\"end\":7.7}]"} {"utterance_id": "3729-6852-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0020.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an ordinal number with fields word, start, end (seconds). If none, reply [].", "labels": [[10.3, 10.73, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":10.3,\"end\":10.73}]"} {"utterance_id": "2830-3980-0073", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0073.flac", "task_type": "lexical_detection", "query": "springs", "prompt": "Process the full audio offline. Find every instance of the word \"springs\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.17, 1.6, "springs"]], "response_mode": "offline", "answer": "[{\"word\":\"springs\",\"start\":1.17,\"end\":1.6}]"} {"utterance_id": "5142-36377-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0024.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "You will hear a complete utterance. When it ends, list every instance of a family relation you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.1, 5.7, "children"]], "response_mode": "offline", "answer": "[{\"word\":\"children\",\"start\":5.1,\"end\":5.7}]"} {"utterance_id": "1284-1181-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0015.flac", "task_type": "lexical_detection", "query": "talks", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"talks\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.75, 4.09, "talks"]], "response_mode": "offline", "answer": "[{\"word\":\"talks\",\"start\":3.75,\"end\":4.09}]"} {"utterance_id": "237-134493-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0012.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.51, 3.14, "cherries"]], "response_mode": "offline", "answer": "[{\"word\":\"cherries\",\"start\":2.51,\"end\":3.14}]"} {"utterance_id": "2300-131720-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0033.flac", "task_type": "semantic_detection", "query": "date", "prompt": "After listening to the whole recording, output every instance of a date with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.6, 2.23, "december eighteen eighty eight"]], "response_mode": "offline", "answer": "[{\"word\":\"december eighteen eighty eight\",\"start\":0.6,\"end\":2.23}]"} {"utterance_id": "4970-29093-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0010.flac", "task_type": "semantic_detection", "query": "city", "prompt": "When the utterance is complete, report all instances of a city name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.54, 3.94, "dorking"]], "response_mode": "offline", "answer": "[{\"word\":\"dorking\",\"start\":3.54,\"end\":3.94}]"} {"utterance_id": "7176-88083-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0013.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "Process the full audio offline. For every occurrence of an animal name, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.01, 2.47, "one"], [12.91, 13.32, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.01,\"end\":2.47},{\"word\":\"two\",\"start\":12.91,\"end\":13.32}]"} {"utterance_id": "7127-75947-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0025.flac", "task_type": "lexical_detection", "query": "exquisite", "prompt": "After listening to the whole recording, output every instance of the word \"exquisite\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.29, 1.07, "exquisite"]], "response_mode": "offline", "answer": "[{\"word\":\"exquisite\",\"start\":0.29,\"end\":1.07}]"} {"utterance_id": "4077-13751-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0009.flac", "task_type": "lexical_counting", "query": "first", "prompt": "When the utterance is complete, report how the word \"first\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.69, 1.06, "one"], [10.12, 10.44, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.69,\"end\":1.06},{\"word\":\"two\",\"start\":10.12,\"end\":10.44}]"} {"utterance_id": "1580-141083-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0043.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.68, 2.53, "miles mc laren"]], "response_mode": "offline", "answer": "[{\"word\":\"miles mc laren\",\"start\":1.68,\"end\":2.53}]"} {"utterance_id": "1320-122617-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0034.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.05, 1.43, "david"]], "response_mode": "offline", "answer": "[{\"word\":\"david\",\"start\":1.05,\"end\":1.43}]"} {"utterance_id": "8555-292519-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0006.flac", "task_type": "lexical_detection", "query": "stroll", "prompt": "After listening to the whole recording, output every instance of the word \"stroll\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.19, 7.68, "stroll"]], "response_mode": "offline", "answer": "[{\"word\":\"stroll\",\"start\":7.19,\"end\":7.68}]"} {"utterance_id": "7127-75947-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0016.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.91, 4.54, "sister"]], "response_mode": "offline", "answer": "[{\"word\":\"sister\",\"start\":3.91,\"end\":4.54}]"} {"utterance_id": "2830-3980-0056", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0056.flac", "task_type": "lexical_detection", "query": "otherwise", "prompt": "Process the full audio offline. Find every instance of the word \"otherwise\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.35, 0.85, "otherwise"]], "response_mode": "offline", "answer": "[{\"word\":\"otherwise\",\"start\":0.35,\"end\":0.85}]"} {"utterance_id": "8463-294825-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0006.flac", "task_type": "lexical_detection", "query": "submarine", "prompt": "When the utterance is complete, report all instances of the word \"submarine\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.39, 3.07, "submarine"]], "response_mode": "offline", "answer": "[{\"word\":\"submarine\",\"start\":2.39,\"end\":3.07}]"} {"utterance_id": "6829-68769-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0013.flac", "task_type": "lexical_detection", "query": "gates", "prompt": "When the utterance is complete, report all instances of the word \"gates\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.92, 1.23, "gates"]], "response_mode": "offline", "answer": "[{\"word\":\"gates\",\"start\":0.92,\"end\":1.23}]"} {"utterance_id": "2300-131720-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0033.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.73, 4.72, "mister w j jenks"]], "response_mode": "offline", "answer": "[{\"word\":\"mister w j jenks\",\"start\":2.73,\"end\":4.72}]"} {"utterance_id": "6829-68771-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0004.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the full audio. After it finishes, find every instance of an organization name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.37, 4.79, "committee"]], "response_mode": "offline", "answer": "[{\"word\":\"committee\",\"start\":4.37,\"end\":4.79}]"} {"utterance_id": "1221-135767-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0021.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.3, 0.68, "pearl"]], "response_mode": "offline", "answer": "[{\"word\":\"pearl\",\"start\":0.3,\"end\":0.68}]"} {"utterance_id": "6829-68771-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0025.flac", "task_type": "lexical_detection", "query": "rocking", "prompt": "When the utterance is complete, report all instances of the word \"rocking\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.2, 1.59, "rocking"]], "response_mode": "offline", "answer": "[{\"word\":\"rocking\",\"start\":1.2,\"end\":1.59}]"} {"utterance_id": "1320-122612-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0008.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[0.43, 0.81, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":0.43,\"end\":0.81}]"} {"utterance_id": "1284-134647-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0004.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[7.72, 8.25, "bishops"]], "response_mode": "offline", "answer": "[{\"word\":\"bishops\",\"start\":7.72,\"end\":8.25}]"} {"utterance_id": "4970-29093-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0011.flac", "task_type": "lexical_detection", "query": "newspaper", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"newspaper\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[7.12, 7.56, "newspaper"]], "response_mode": "offline", "answer": "[{\"word\":\"newspaper\",\"start\":7.12,\"end\":7.56}]"} {"utterance_id": "4970-29095-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0036.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "After listening to the whole recording, output timed count words for a profession (one, two, three, ...) as JSON. If none, reply [].", "labels": [[2.04, 2.53, "one"], [3.67, 4.06, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.04,\"end\":2.53},{\"word\":\"two\",\"start\":3.67,\"end\":4.06}]"} {"utterance_id": "6829-68769-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0025.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.49, 0.92, "rogers"]], "response_mode": "offline", "answer": "[{\"word\":\"rogers\",\"start\":0.49,\"end\":0.92}]"} {"utterance_id": "6829-68771-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0036.flac", "task_type": "lexical_detection", "query": "decided", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"decided\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.13, 2.69, "decided"]], "response_mode": "offline", "answer": "[{\"word\":\"decided\",\"start\":2.13,\"end\":2.69}]"} {"utterance_id": "1320-122617-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0028.flac", "task_type": "lexical_detection", "query": "pursuits", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"pursuits\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.69, 1.22, "pursuits"]], "response_mode": "offline", "answer": "[{\"word\":\"pursuits\",\"start\":0.69,\"end\":1.22}]"} {"utterance_id": "2094-142345-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0038.flac", "task_type": "lexical_detection", "query": "souls", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"souls\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.37, 4.7, "souls"]], "response_mode": "offline", "answer": "[{\"word\":\"souls\",\"start\":4.37,\"end\":4.7}]"} {"utterance_id": "672-122797-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0041.flac", "task_type": "lexical_detection", "query": "servant", "prompt": "Once the audio has finished, identify every instance of the word \"servant\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.43, 2.09, "servant"]], "response_mode": "offline", "answer": "[{\"word\":\"servant\",\"start\":1.43,\"end\":2.09}]"} {"utterance_id": "4992-41797-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0014.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.38, 4.81, "father's"]], "response_mode": "offline", "answer": "[{\"word\":\"father's\",\"start\":4.38,\"end\":4.81}]"} {"utterance_id": "1320-122612-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0007.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.24, 3.74, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":3.24,\"end\":3.74}]"} {"utterance_id": "4446-2275-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.83, 1.24, "bartley"]], "response_mode": "offline", "answer": "[{\"word\":\"bartley\",\"start\":0.83,\"end\":1.24}]"} {"utterance_id": "908-157963-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0029.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.83, 1.18, "tongue"]], "response_mode": "offline", "answer": "[{\"word\":\"tongue\",\"start\":0.83,\"end\":1.18}]"} {"utterance_id": "2961-960-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0011.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After the audio ends, extract all occurrences of an organization name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[15.35, 16.39, "pythagoreans"]], "response_mode": "offline", "answer": "[{\"word\":\"pythagoreans\",\"start\":15.35,\"end\":16.39}]"} {"utterance_id": "2094-142345-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0037.flac", "task_type": "lexical_detection", "query": "wherever", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"wherever\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.0, 2.48, "wherever"]], "response_mode": "offline", "answer": "[{\"word\":\"wherever\",\"start\":2.0,\"end\":2.48}]"} {"utterance_id": "237-134500-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0012.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[0.5, 0.73, "few"], [8.63, 8.91, "few"]], "response_mode": "offline", "answer": "[{\"word\":\"few\",\"start\":0.5,\"end\":0.73},{\"word\":\"few\",\"start\":8.63,\"end\":8.91}]"} {"utterance_id": "1284-1181-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.64, 6.08, "glinda"]], "response_mode": "offline", "answer": "[{\"word\":\"glinda\",\"start\":5.64,\"end\":6.08}]"} {"utterance_id": "1188-133604-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0017.flac", "task_type": "lexical_detection", "query": "erroneous", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"erroneous\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.76, 4.45, "erroneous"]], "response_mode": "offline", "answer": "[{\"word\":\"erroneous\",\"start\":3.76,\"end\":4.45}]"} {"utterance_id": "8555-284447-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0021.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Do not answer until the audio ends. Then list each instance of an animal name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.98, 3.21, "goat"]], "response_mode": "offline", "answer": "[{\"word\":\"goat\",\"start\":2.98,\"end\":3.21}]"} {"utterance_id": "2094-142345-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0045.flac", "task_type": "lexical_detection", "query": "poyser", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"poyser\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.34, 2.78, "poyser"]], "response_mode": "offline", "answer": "[{\"word\":\"poyser\",\"start\":2.34,\"end\":2.78}]"} {"utterance_id": "8555-284447-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0015.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[4.8, 5.04, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.8,\"end\":5.04}]"} {"utterance_id": "3570-5695-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0000.flac", "task_type": "lexical_detection", "query": "groups", "prompt": "After listening to the whole recording, output every instance of the word \"groups\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.3, 3.7, "groups"]], "response_mode": "offline", "answer": "[{\"word\":\"groups\",\"start\":3.3,\"end\":3.7}]"} {"utterance_id": "1580-141084-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0041.flac", "task_type": "lexical_detection", "query": "harm", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"harm\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.57, 0.86, "harm"]], "response_mode": "offline", "answer": "[{\"word\":\"harm\",\"start\":0.57,\"end\":0.86}]"} {"utterance_id": "1995-1837-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0023.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Listen to the entire clip, then report the count sequence for a body part as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.96, 12.46, "one"], [15.6, 16.02, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":11.96,\"end\":12.46},{\"word\":\"two\",\"start\":15.6,\"end\":16.02}]"} {"utterance_id": "1320-122617-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.53, 2.9, "david"]], "response_mode": "offline", "answer": "[{\"word\":\"david\",\"start\":2.53,\"end\":2.9}]"} {"utterance_id": "2300-131720-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0018.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen to the full audio. After it finishes, find every instance of a money or currency term. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.75, 3.47, "economy"]], "response_mode": "offline", "answer": "[{\"word\":\"economy\",\"start\":2.75,\"end\":3.47}]"} {"utterance_id": "672-122797-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0050.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a plant name: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[1.6, 2.39, "one"], [3.68, 4.44, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.6,\"end\":2.39},{\"word\":\"two\",\"start\":3.68,\"end\":4.44}]"} {"utterance_id": "672-122797-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0042.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Once the audio has finished, identify every instance of a color name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[9.43, 9.82, "dark"]], "response_mode": "offline", "answer": "[{\"word\":\"dark\",\"start\":9.43,\"end\":9.82}]"} {"utterance_id": "8463-294825-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0002.flac", "task_type": "lexical_detection", "query": "frequent", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"frequent\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.11, 6.57, "frequent"]], "response_mode": "offline", "answer": "[{\"word\":\"frequent\",\"start\":6.11,\"end\":6.57}]"} {"utterance_id": "908-31957-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0014.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.24, 6.73, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":6.24,\"end\":6.73}]"} {"utterance_id": "7729-102255-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0003.flac", "task_type": "lexical_detection", "query": "special", "prompt": "When the utterance is complete, report all instances of the word \"special\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.22, 3.8, "special"]], "response_mode": "offline", "answer": "[{\"word\":\"special\",\"start\":3.22,\"end\":3.8}]"} {"utterance_id": "3729-6852-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0036.flac", "task_type": "lexical_detection", "query": "paris", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"paris\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.32, 1.94, "paris"]], "response_mode": "offline", "answer": "[{\"word\":\"paris\",\"start\":1.32,\"end\":1.94}]"} {"utterance_id": "61-70968-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0006.flac", "task_type": "lexical_detection", "query": "quickly", "prompt": "Once the audio has finished, identify every instance of the word \"quickly\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.65, 2.02, "quickly"]], "response_mode": "offline", "answer": "[{\"word\":\"quickly\",\"start\":1.65,\"end\":2.02}]"} {"utterance_id": "4992-23283-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0005.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.74, 2.02, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.74,\"end\":2.02}]"} {"utterance_id": "4970-29093-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0010.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, list every instance of an organization name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.91, 11.92, "daily grapevine"]], "response_mode": "offline", "answer": "[{\"word\":\"daily grapevine\",\"start\":10.91,\"end\":11.92}]"} {"utterance_id": "4446-2271-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0024.flac", "task_type": "lexical_detection", "query": "laugh", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"laugh\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.58, 1.89, "laugh"]], "response_mode": "offline", "answer": "[{\"word\":\"laugh\",\"start\":1.58,\"end\":1.89}]"} {"utterance_id": "5142-36377-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0024.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.79, 1.31, "forefinger"]], "response_mode": "offline", "answer": "[{\"word\":\"forefinger\",\"start\":0.79,\"end\":1.31}]"} {"utterance_id": "237-126133-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0008.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Once the audio has finished, list each occurrence of a person's name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[0.54, 1.12, "one"], [2.57, 3.05, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.54,\"end\":1.12},{\"word\":\"two\",\"start\":2.57,\"end\":3.05}]"} {"utterance_id": "8463-294828-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0031.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.7, 1.21, "sailors"], [5.19, 5.77, "officer"]], "response_mode": "offline", "answer": "[{\"word\":\"sailors\",\"start\":0.7,\"end\":1.21},{\"word\":\"officer\",\"start\":5.19,\"end\":5.77}]"} {"utterance_id": "4507-16021-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0048.flac", "task_type": "lexical_detection", "query": "goes", "prompt": "Once the audio has finished, identify every instance of the word \"goes\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.94, 4.27, "goes"]], "response_mode": "offline", "answer": "[{\"word\":\"goes\",\"start\":3.94,\"end\":4.27}]"} {"utterance_id": "6829-68771-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0031.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "After the audio ends, tally every instance of a body part. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.45, 0.71, "one"], [1.81, 2.23, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.45,\"end\":0.71},{\"word\":\"two\",\"start\":1.81,\"end\":2.23}]"} {"utterance_id": "4077-13754-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0008.flac", "task_type": "lexical_detection", "query": "nations", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"nations\" with fields word, start, end (seconds). If none, reply [].", "labels": [[7.0, 7.4, "nations"]], "response_mode": "offline", "answer": "[{\"word\":\"nations\",\"start\":7.0,\"end\":7.4}]"} {"utterance_id": "5683-32865-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0015.flac", "task_type": "lexical_detection", "query": "night", "prompt": "Process the full audio offline. Find every instance of the word \"night\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.2, 2.54, "night"]], "response_mode": "offline", "answer": "[{\"word\":\"night\",\"start\":2.2,\"end\":2.54}]"} {"utterance_id": "6930-75918-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0000.flac", "task_type": "semantic_detection", "query": "city", "prompt": "When the utterance is complete, report all instances of a city name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 1.19, "concord"]], "response_mode": "offline", "answer": "[{\"word\":\"concord\",\"start\":0.53,\"end\":1.19}]"} {"utterance_id": "7021-79740-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0011.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen carefully to the entire clip, then report every instance of a family relation as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.27, 4.85, "children"]], "response_mode": "offline", "answer": "[{\"word\":\"children\",\"start\":4.27,\"end\":4.85}]"} {"utterance_id": "8555-292519-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0011.flac", "task_type": "lexical_detection", "query": "courtyard", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"courtyard\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.91, 2.74, "courtyard"]], "response_mode": "offline", "answer": "[{\"word\":\"courtyard\",\"start\":1.91,\"end\":2.74}]"} {"utterance_id": "1995-1826-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0003.flac", "task_type": "lexical_detection", "query": "better", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"better\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.09, 0.59, "better"]], "response_mode": "offline", "answer": "[{\"word\":\"better\",\"start\":0.09,\"end\":0.59}]"} {"utterance_id": "1089-134686-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0034.flac", "task_type": "lexical_detection", "query": "shaking", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"shaking\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.36, 2.84, "shaking"]], "response_mode": "offline", "answer": "[{\"word\":\"shaking\",\"start\":2.36,\"end\":2.84}]"} {"utterance_id": "1188-133604-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0041.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.28, 2.56, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.28,\"end\":2.56}]"} {"utterance_id": "1320-122617-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0041.flac", "task_type": "lexical_detection", "query": "proportions", "prompt": "After listening to the whole recording, output every instance of the word \"proportions\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.28, 4.05, "proportions"]], "response_mode": "offline", "answer": "[{\"word\":\"proportions\",\"start\":3.28,\"end\":4.05}]"} {"utterance_id": "8224-274384-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0005.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.95, 1.62, "preacher"]], "response_mode": "offline", "answer": "[{\"word\":\"preacher\",\"start\":0.95,\"end\":1.62}]"} {"utterance_id": "1089-134686-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0033.flac", "task_type": "lexical_detection", "query": "francis", "prompt": "After listening to the whole recording, output every instance of the word \"francis\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.91, 2.48, "francis"]], "response_mode": "offline", "answer": "[{\"word\":\"francis\",\"start\":1.91,\"end\":2.48}]"} {"utterance_id": "4507-16021-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0025.flac", "task_type": "lexical_detection", "query": "needs", "prompt": "When the utterance is complete, report all instances of the word \"needs\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.28, 1.66, "needs"]], "response_mode": "offline", "answer": "[{\"word\":\"needs\",\"start\":1.28,\"end\":1.66}]"} {"utterance_id": "5683-32879-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0002.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the full audio. After it finishes, find every instance of a color name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.17, 9.55, "pale"]], "response_mode": "offline", "answer": "[{\"word\":\"pale\",\"start\":9.17,\"end\":9.55}]"} {"utterance_id": "2830-3980-0062", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0062.flac", "task_type": "lexical_detection", "query": "kingdom", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"kingdom\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.93, 2.5, "kingdom"]], "response_mode": "offline", "answer": "[{\"word\":\"kingdom\",\"start\":1.93,\"end\":2.5}]"} {"utterance_id": "2830-3980-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0010.flac", "task_type": "lexical_detection", "query": "directly", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"directly\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.95, 4.4, "directly"]], "response_mode": "offline", "answer": "[{\"word\":\"directly\",\"start\":3.95,\"end\":4.4}]"} {"utterance_id": "7021-79730-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0003.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.75, 12.31, "doctor"]], "response_mode": "offline", "answer": "[{\"word\":\"doctor\",\"start\":11.75,\"end\":12.31}]"} {"utterance_id": "3570-5694-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0001.flac", "task_type": "lexical_detection", "query": "utility", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"utility\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.6, 1.08, "utility"]], "response_mode": "offline", "answer": "[{\"word\":\"utility\",\"start\":0.6,\"end\":1.08}]"} {"utterance_id": "1284-1180-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0017.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "After the audio ends, tally every instance of a number. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.32, 0.56, "one"], [6.62, 6.99, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.32,\"end\":0.56},{\"word\":\"two\",\"start\":6.62,\"end\":6.99}]"} {"utterance_id": "3729-6852-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0023.flac", "task_type": "lexical_detection", "query": "writer", "prompt": "When the utterance is complete, report all instances of the word \"writer\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.39, 7.8, "writer"]], "response_mode": "offline", "answer": "[{\"word\":\"writer\",\"start\":7.39,\"end\":7.8}]"} {"utterance_id": "4970-29095-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0017.flac", "task_type": "lexical_detection", "query": "spirits", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"spirits\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.32, 2.88, "spirits"]], "response_mode": "offline", "answer": "[{\"word\":\"spirits\",\"start\":2.32,\"end\":2.88}]"} {"utterance_id": "7127-75947-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0026.flac", "task_type": "lexical_detection", "query": "charente", "prompt": "After listening to the whole recording, output every instance of the word \"charente\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.25, 2.76, "charente"]], "response_mode": "offline", "answer": "[{\"word\":\"charente\",\"start\":2.25,\"end\":2.76}]"} {"utterance_id": "2094-142345-0057", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0057.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a number with word/start/end. If none, reply [].", "labels": [[3.17, 3.41, "one"], [8.42, 8.64, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.17,\"end\":3.41},{\"word\":\"two\",\"start\":8.42,\"end\":8.64}]"} {"utterance_id": "7176-88083-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0002.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Do not answer until the audio ends. Then list each instance of a color name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.96, 1.51, "red"], [6.86, 7.24, "red"]], "response_mode": "offline", "answer": "[{\"word\":\"red\",\"start\":0.96,\"end\":1.51},{\"word\":\"red\",\"start\":6.86,\"end\":7.24}]"} {"utterance_id": "6829-68769-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0005.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.91, 3.13, "beth"]], "response_mode": "offline", "answer": "[{\"word\":\"beth\",\"start\":2.91,\"end\":3.13}]"} {"utterance_id": "3575-170457-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0046.flac", "task_type": "lexical_detection", "query": "neighbour", "prompt": "When the utterance is complete, report all instances of the word \"neighbour\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.85, 1.17, "neighbour"]], "response_mode": "offline", "answer": "[{\"word\":\"neighbour\",\"start\":0.85,\"end\":1.17}]"} {"utterance_id": "1580-141083-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0040.flac", "task_type": "lexical_detection", "query": "throw", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"throw\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.3, 1.5, "throw"]], "response_mode": "offline", "answer": "[{\"word\":\"throw\",\"start\":1.3,\"end\":1.5}]"} {"utterance_id": "1284-1181-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0020.flac", "task_type": "lexical_detection", "query": "pleased", "prompt": "When the utterance is complete, report all instances of the word \"pleased\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.23, 5.7, "pleased"]], "response_mode": "offline", "answer": "[{\"word\":\"pleased\",\"start\":5.23,\"end\":5.7}]"} {"utterance_id": "3570-5696-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0009.flac", "task_type": "semantic_detection", "query": "money", "prompt": "After listening to the whole recording, output every instance of a money or currency term with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.65, 5.31, "expenditure"], [7.19, 7.78, "pecuniary"]], "response_mode": "offline", "answer": "[{\"word\":\"expenditure\",\"start\":4.65,\"end\":5.31},{\"word\":\"pecuniary\",\"start\":7.19,\"end\":7.78}]"} {"utterance_id": "1284-1181-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0016.flac", "task_type": "lexical_detection", "query": "pipe", "prompt": "Once the audio has finished, identify every instance of the word \"pipe\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.23, 6.69, "pipe"]], "response_mode": "offline", "answer": "[{\"word\":\"pipe\",\"start\":6.23,\"end\":6.69}]"} {"utterance_id": "7021-85628-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.65, 6.07, "anders"]], "response_mode": "offline", "answer": "[{\"word\":\"anders\",\"start\":5.65,\"end\":6.07}]"} {"utterance_id": "8555-284447-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0015.flac", "task_type": "lexical_detection", "query": "suggested", "prompt": "After the audio ends, extract all occurrences of the word \"suggested\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.5, 3.1, "suggested"]], "response_mode": "offline", "answer": "[{\"word\":\"suggested\",\"start\":2.5,\"end\":3.1}]"} {"utterance_id": "2094-142345-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0024.flac", "task_type": "lexical_detection", "query": "warm", "prompt": "When the utterance is complete, report all instances of the word \"warm\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.25, 4.83, "warm"]], "response_mode": "offline", "answer": "[{\"word\":\"warm\",\"start\":4.25,\"end\":4.83}]"} {"utterance_id": "1320-122612-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0012.flac", "task_type": "lexical_detection", "query": "brands", "prompt": "Process the full audio offline. Find every instance of the word \"brands\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.03, 1.51, "brands"]], "response_mode": "offline", "answer": "[{\"word\":\"brands\",\"start\":1.03,\"end\":1.51}]"} {"utterance_id": "1221-135767-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0007.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.17, 2.56, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.17,\"end\":2.56}]"} {"utterance_id": "7176-92135-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0043.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a food name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.47, 1.72, "bread"]], "response_mode": "offline", "answer": "[{\"word\":\"bread\",\"start\":1.47,\"end\":1.72}]"} {"utterance_id": "2094-142345-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0046.flac", "task_type": "lexical_detection", "query": "delight", "prompt": "Process the full audio offline. Find every instance of the word \"delight\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.63, 0.97, "delight"]], "response_mode": "offline", "answer": "[{\"word\":\"delight\",\"start\":0.63,\"end\":0.97}]"} {"utterance_id": "1320-122617-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0022.flac", "task_type": "lexical_detection", "query": "deer", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"deer\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.86, 3.34, "deer"]], "response_mode": "offline", "answer": "[{\"word\":\"deer\",\"start\":2.86,\"end\":3.34}]"} {"utterance_id": "260-123288-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0015.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Listen carefully to the entire clip, then report every instance of a weather term as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.59, 2.22, "clouds"]], "response_mode": "offline", "answer": "[{\"word\":\"clouds\",\"start\":1.59,\"end\":2.22}]"} {"utterance_id": "121-127105-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0011.flac", "task_type": "lexical_detection", "query": "agreeable", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"agreeable\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.0, 1.57, "agreeable"]], "response_mode": "offline", "answer": "[{\"word\":\"agreeable\",\"start\":1.0,\"end\":1.57}]"} {"utterance_id": "121-127105-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0027.flac", "task_type": "lexical_detection", "query": "spoils", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"spoils\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.47, 3.99, "spoils"]], "response_mode": "offline", "answer": "[{\"word\":\"spoils\",\"start\":3.47,\"end\":3.99}]"} {"utterance_id": "2961-960-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0015.flac", "task_type": "lexical_detection", "query": "dialogues", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"dialogues\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.66, 7.39, "dialogues"]], "response_mode": "offline", "answer": "[{\"word\":\"dialogues\",\"start\":6.66,\"end\":7.39}]"} {"utterance_id": "1320-122617-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0040.flac", "task_type": "lexical_detection", "query": "delay", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"delay\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.52, 1.92, "delay"]], "response_mode": "offline", "answer": "[{\"word\":\"delay\",\"start\":1.52,\"end\":1.92}]"} {"utterance_id": "121-121726-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0000.flac", "task_type": "lexical_detection", "query": "contrivance", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"contrivance\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.62, 2.32, "contrivance"]], "response_mode": "offline", "answer": "[{\"word\":\"contrivance\",\"start\":1.62,\"end\":2.32}]"} {"utterance_id": "5639-40744-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0040.flac", "task_type": "lexical_detection", "query": "sufficient", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"sufficient\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.11, 5.68, "sufficient"]], "response_mode": "offline", "answer": "[{\"word\":\"sufficient\",\"start\":5.11,\"end\":5.68}]"} {"utterance_id": "4970-29095-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0001.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.3, 2.68, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":2.3,\"end\":2.68}]"} {"utterance_id": "7729-102255-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0007.flac", "task_type": "lexical_detection", "query": "westport", "prompt": "After listening to the whole recording, output every instance of the word \"westport\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.59, 2.26, "westport"]], "response_mode": "offline", "answer": "[{\"word\":\"westport\",\"start\":1.59,\"end\":2.26}]"} {"utterance_id": "4446-2275-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0013.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.88, 2.3, "bartley"]], "response_mode": "offline", "answer": "[{\"word\":\"bartley\",\"start\":1.88,\"end\":2.3}]"} {"utterance_id": "2961-960-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0005.flac", "task_type": "lexical_detection", "query": "ancient", "prompt": "Process the full audio offline. Find every instance of the word \"ancient\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.83, 4.27, "ancient"]], "response_mode": "offline", "answer": "[{\"word\":\"ancient\",\"start\":3.83,\"end\":4.27}]"} {"utterance_id": "2094-142345-0055", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0055.flac", "task_type": "lexical_detection", "query": "sir", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"sir\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.02, 2.44, "sir"]], "response_mode": "offline", "answer": "[{\"word\":\"sir\",\"start\":2.02,\"end\":2.44}]"} {"utterance_id": "3729-6852-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0005.flac", "task_type": "lexical_counting", "query": "comedies", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"comedies\" with word/start/end. If none, reply [].", "labels": [[6.58, 7.42, "one"], [15.31, 15.92, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.58,\"end\":7.42},{\"word\":\"two\",\"start\":15.31,\"end\":15.92}]"} {"utterance_id": "61-70968-0062", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0062.flac", "task_type": "lexical_detection", "query": "tricks", "prompt": "Process the full audio offline. Find every instance of the word \"tricks\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.57, 2.13, "tricks"]], "response_mode": "offline", "answer": "[{\"word\":\"tricks\",\"start\":1.57,\"end\":2.13}]"} {"utterance_id": "3729-6852-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0005.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "After the audio ends, tally every instance of a person's name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.49, 1.08, "one"], [12.18, 13.03, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.49,\"end\":1.08},{\"word\":\"two\",\"start\":12.18,\"end\":13.03}]"} {"utterance_id": "4992-41797-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0009.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After listening to the whole recording, output every instance of an organization name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.95, 7.6, "college"]], "response_mode": "offline", "answer": "[{\"word\":\"college\",\"start\":6.95,\"end\":7.6}]"} {"utterance_id": "2961-961-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0016.flac", "task_type": "lexical_detection", "query": "describe", "prompt": "After the audio ends, extract all occurrences of the word \"describe\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.33, 1.93, "describe"]], "response_mode": "offline", "answer": "[{\"word\":\"describe\",\"start\":1.33,\"end\":1.93}]"} {"utterance_id": "1284-1180-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0000.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a body part with word/start/end. If none, reply [].", "labels": [[2.37, 2.58, "one"], [5.02, 5.46, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.37,\"end\":2.58},{\"word\":\"two\",\"start\":5.02,\"end\":5.46}]"} {"utterance_id": "8463-294825-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0012.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.29, 3.62, "nemo"]], "response_mode": "offline", "answer": "[{\"word\":\"nemo\",\"start\":3.29,\"end\":3.62}]"} {"utterance_id": "8463-294828-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0002.flac", "task_type": "lexical_detection", "query": "badly", "prompt": "After the audio ends, extract all occurrences of the word \"badly\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.64, 5.01, "badly"]], "response_mode": "offline", "answer": "[{\"word\":\"badly\",\"start\":4.64,\"end\":5.01}]"} {"utterance_id": "8224-274384-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0002.flac", "task_type": "lexical_detection", "query": "informed", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"informed\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.72, 1.51, "informed"]], "response_mode": "offline", "answer": "[{\"word\":\"informed\",\"start\":0.72,\"end\":1.51}]"} {"utterance_id": "5639-40744-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0039.flac", "task_type": "lexical_detection", "query": "senses", "prompt": "Process the full audio offline. Find every instance of the word \"senses\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.05, 1.68, "senses"]], "response_mode": "offline", "answer": "[{\"word\":\"senses\",\"start\":1.05,\"end\":1.68}]"} {"utterance_id": "237-134493-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0016.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After the audio ends, extract all occurrences of a plant name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.43, 6.93, "osage orange hedges"]], "response_mode": "offline", "answer": "[{\"word\":\"osage orange hedges\",\"start\":5.43,\"end\":6.93}]"} {"utterance_id": "8224-274381-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0002.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.67, 1.36, "former"], [14.72, 15.26, "latter"]], "response_mode": "offline", "answer": "[{\"word\":\"former\",\"start\":0.67,\"end\":1.36},{\"word\":\"latter\",\"start\":14.72,\"end\":15.26}]"} {"utterance_id": "7127-75947-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0018.flac", "task_type": "lexical_detection", "query": "valliere", "prompt": "After the audio ends, extract all occurrences of the word \"valliere\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.16, 3.65, "valliere"]], "response_mode": "offline", "answer": "[{\"word\":\"valliere\",\"start\":3.16,\"end\":3.65}]"} {"utterance_id": "1188-133604-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0026.flac", "task_type": "lexical_counting", "query": "mist", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"mist\" with word/start/end. If none, reply [].", "labels": [[5.39, 5.86, "one"], [17.01, 17.5, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.39,\"end\":5.86},{\"word\":\"two\",\"start\":17.01,\"end\":17.5}]"} {"utterance_id": "6829-68769-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0008.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen carefully to the entire clip, then report every instance of a time of day as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.15, 2.72, "ten o'clock"]], "response_mode": "offline", "answer": "[{\"word\":\"ten o'clock\",\"start\":2.15,\"end\":2.72}]"} {"utterance_id": "2830-3980-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.65, 1.88, "paul"]], "response_mode": "offline", "answer": "[{\"word\":\"paul\",\"start\":1.65,\"end\":1.88}]"} {"utterance_id": "7127-75947-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0001.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.99, 3.5, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":2.99,\"end\":3.5}]"} {"utterance_id": "3575-170457-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[11.29, 12.19, "robert southey"]], "response_mode": "offline", "answer": "[{\"word\":\"robert southey\",\"start\":11.29,\"end\":12.19}]"} {"utterance_id": "2094-142345-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0006.flac", "task_type": "lexical_detection", "query": "resemblance", "prompt": "When the utterance is complete, report all instances of the word \"resemblance\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.5, 9.22, "resemblance"]], "response_mode": "offline", "answer": "[{\"word\":\"resemblance\",\"start\":8.5,\"end\":9.22}]"} {"utterance_id": "237-134500-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0011.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Once the audio has finished, identify every instance of a plant name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.71, 3.22, "grass"]], "response_mode": "offline", "answer": "[{\"word\":\"grass\",\"start\":2.71,\"end\":3.22}]"} {"utterance_id": "237-126133-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0000.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "After the audio ends, tally every instance of a person's name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[6.54, 6.89, "one"], [8.04, 8.45, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.54,\"end\":6.89},{\"word\":\"two\",\"start\":8.04,\"end\":8.45}]"} {"utterance_id": "1580-141084-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0037.flac", "task_type": "lexical_detection", "query": "approached", "prompt": "Process the full audio offline. Find every instance of the word \"approached\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.63, 1.07, "approached"]], "response_mode": "offline", "answer": "[{\"word\":\"approached\",\"start\":0.63,\"end\":1.07}]"} {"utterance_id": "7176-92135-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0045.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Once the audio has finished, identify every instance of a color name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.72, 2.03, "gold"]], "response_mode": "offline", "answer": "[{\"word\":\"gold\",\"start\":1.72,\"end\":2.03}]"} {"utterance_id": "5105-28241-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0010.flac", "task_type": "lexical_detection", "query": "supreme", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"supreme\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.29, 1.94, "supreme"]], "response_mode": "offline", "answer": "[{\"word\":\"supreme\",\"start\":1.29,\"end\":1.94}]"} {"utterance_id": "5639-40744-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0038.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[12.84, 13.19, "hearts"]], "response_mode": "offline", "answer": "[{\"word\":\"hearts\",\"start\":12.84,\"end\":13.19}]"} {"utterance_id": "1580-141084-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0003.flac", "task_type": "lexical_detection", "query": "holmes", "prompt": "After listening to the whole recording, output every instance of the word \"holmes\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.83, 2.3, "holmes"]], "response_mode": "offline", "answer": "[{\"word\":\"holmes\",\"start\":1.83,\"end\":2.3}]"} {"utterance_id": "4992-23283-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0017.flac", "task_type": "lexical_detection", "query": "breaking", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"breaking\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[8.07, 8.52, "breaking"]], "response_mode": "offline", "answer": "[{\"word\":\"breaking\",\"start\":8.07,\"end\":8.52}]"} {"utterance_id": "4077-13754-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0011.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a country name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.45, 2.07, "united states"]], "response_mode": "offline", "answer": "[{\"word\":\"united states\",\"start\":1.45,\"end\":2.07}]"} {"utterance_id": "8463-287645-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0007.flac", "task_type": "lexical_detection", "query": "rage", "prompt": "After listening to the whole recording, output every instance of the word \"rage\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.66, 13.07, "rage"]], "response_mode": "offline", "answer": "[{\"word\":\"rage\",\"start\":12.66,\"end\":13.07}]"} {"utterance_id": "6930-75918-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0014.flac", "task_type": "lexical_detection", "query": "pronounced", "prompt": "Once the audio has finished, identify every instance of the word \"pronounced\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[7.34, 8.0, "pronounced"]], "response_mode": "offline", "answer": "[{\"word\":\"pronounced\",\"start\":7.34,\"end\":8.0}]"} {"utterance_id": "3570-5694-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0010.flac", "task_type": "lexical_detection", "query": "paraphernalia", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"paraphernalia\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.76, 5.64, "paraphernalia"]], "response_mode": "offline", "answer": "[{\"word\":\"paraphernalia\",\"start\":4.76,\"end\":5.64}]"} {"utterance_id": "8463-294828-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0026.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.56, 1.14, "commander"]], "response_mode": "offline", "answer": "[{\"word\":\"commander\",\"start\":0.56,\"end\":1.14}]"} {"utterance_id": "1221-135767-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0014.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.66, 4.25, "minister"], [6.15, 6.74, "leech"]], "response_mode": "offline", "answer": "[{\"word\":\"minister\",\"start\":3.66,\"end\":4.25},{\"word\":\"leech\",\"start\":6.15,\"end\":6.74}]"} {"utterance_id": "61-70968-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0003.flac", "task_type": "semantic_counting", "query": "kinship", "prompt": "After the audio ends, tally every instance of a family relation. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[1.15, 1.66, "one"], [3.47, 3.98, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.15,\"end\":1.66},{\"word\":\"two\",\"start\":3.47,\"end\":3.98}]"} {"utterance_id": "1995-1826-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0020.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.92, 4.08, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.92,\"end\":4.08}]"} {"utterance_id": "8463-294828-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0024.flac", "task_type": "lexical_detection", "query": "take", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"take\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.19, 1.54, "take"]], "response_mode": "offline", "answer": "[{\"word\":\"take\",\"start\":1.19,\"end\":1.54}]"} {"utterance_id": "1284-1180-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0031.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.77, 2.7, "princess ozma"]], "response_mode": "offline", "answer": "[{\"word\":\"princess ozma\",\"start\":1.77,\"end\":2.7}]"} {"utterance_id": "8230-279154-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0001.flac", "task_type": "lexical_detection", "query": "sensation", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"sensation\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.9, 3.78, "sensation"], [16.32, 17.12, "sensation"]], "response_mode": "offline", "answer": "[{\"word\":\"sensation\",\"start\":2.9,\"end\":3.78},{\"word\":\"sensation\",\"start\":16.32,\"end\":17.12}]"} {"utterance_id": "5142-33396-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0054.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.2, 4.4, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.2,\"end\":4.4}]"} {"utterance_id": "4992-41797-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.43, 1.24, "henry lord"]], "response_mode": "offline", "answer": "[{\"word\":\"henry lord\",\"start\":0.43,\"end\":1.24}]"} {"utterance_id": "4077-13751-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0008.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Process the full audio offline. Find every instance of an organization name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.95, 4.86, "latter day saints"]], "response_mode": "offline", "answer": "[{\"word\":\"latter day saints\",\"start\":3.95,\"end\":4.86}]"} {"utterance_id": "7729-102255-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0011.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Do not answer until the audio ends. Then list each instance of a state or province name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.68, 5.31, "kansas"]], "response_mode": "offline", "answer": "[{\"word\":\"kansas\",\"start\":4.68,\"end\":5.31}]"} {"utterance_id": "4992-41797-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0000.flac", "task_type": "lexical_detection", "query": "job", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"job\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.83, 4.19, "job"]], "response_mode": "offline", "answer": "[{\"word\":\"job\",\"start\":3.83,\"end\":4.19}]"} {"utterance_id": "4077-13751-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0006.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Process the full audio offline. Find every instance of a state or province name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.73, 9.2, "west"]], "response_mode": "offline", "answer": "[{\"word\":\"west\",\"start\":8.73,\"end\":9.2}]"} {"utterance_id": "7176-92135-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0021.flac", "task_type": "lexical_detection", "query": "wondering", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"wondering\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.0, 1.42, "wondering"]], "response_mode": "offline", "answer": "[{\"word\":\"wondering\",\"start\":1.0,\"end\":1.42}]"} {"utterance_id": "1320-122617-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0016.flac", "task_type": "lexical_detection", "query": "blow", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"blow\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.45, 2.7, "blow"]], "response_mode": "offline", "answer": "[{\"word\":\"blow\",\"start\":2.45,\"end\":2.7}]"} {"utterance_id": "237-134500-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0040.flac", "task_type": "lexical_detection", "query": "pray", "prompt": "When the utterance is complete, report all instances of the word \"pray\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.39, 0.62, "pray"]], "response_mode": "offline", "answer": "[{\"word\":\"pray\",\"start\":0.39,\"end\":0.62}]"} {"utterance_id": "5639-40744-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0007.flac", "task_type": "lexical_detection", "query": "custody", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"custody\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.21, 3.75, "custody"]], "response_mode": "offline", "answer": "[{\"word\":\"custody\",\"start\":3.21,\"end\":3.75}]"} {"utterance_id": "1188-133604-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0008.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the full audio. After it finishes, find every instance of an ordinal number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.14, 2.46, "final"]], "response_mode": "offline", "answer": "[{\"word\":\"final\",\"start\":2.14,\"end\":2.46}]"} {"utterance_id": "61-70968-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0004.flac", "task_type": "lexical_detection", "query": "page", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"page\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.82, 2.31, "page"]], "response_mode": "offline", "answer": "[{\"word\":\"page\",\"start\":1.82,\"end\":2.31}]"} {"utterance_id": "5683-32866-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.86, 6.19, "rachel"]], "response_mode": "offline", "answer": "[{\"word\":\"rachel\",\"start\":5.86,\"end\":6.19}]"} {"utterance_id": "237-134500-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0037.flac", "task_type": "lexical_detection", "query": "nice", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"nice\" with fields word, start, end (seconds). If none, reply [].", "labels": [[5.17, 5.53, "nice"]], "response_mode": "offline", "answer": "[{\"word\":\"nice\",\"start\":5.17,\"end\":5.53}]"} {"utterance_id": "260-123286-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0030.flac", "task_type": "lexical_detection", "query": "disappear", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"disappear\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.73, 4.3, "disappear"]], "response_mode": "offline", "answer": "[{\"word\":\"disappear\",\"start\":3.73,\"end\":4.3}]"} {"utterance_id": "6829-68769-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0018.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.95, 3.24, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":2.95,\"end\":3.24}]"} {"utterance_id": "6930-75918-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.69, 2.06, "raoul"]], "response_mode": "offline", "answer": "[{\"word\":\"raoul\",\"start\":1.69,\"end\":2.06}]"} {"utterance_id": "237-134500-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0032.flac", "task_type": "lexical_detection", "query": "tired", "prompt": "Once the audio has finished, identify every instance of the word \"tired\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.48, 0.84, "tired"]], "response_mode": "offline", "answer": "[{\"word\":\"tired\",\"start\":0.48,\"end\":0.84}]"} {"utterance_id": "908-31957-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0022.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "When the utterance is complete, report all instances of a time of day including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.53, 11.93, "morning"]], "response_mode": "offline", "answer": "[{\"word\":\"morning\",\"start\":11.53,\"end\":11.93}]"} {"utterance_id": "3575-170457-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0038.flac", "task_type": "lexical_detection", "query": "meantime", "prompt": "Process the full audio offline. Find every instance of the word \"meantime\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.62, 1.24, "meantime"]], "response_mode": "offline", "answer": "[{\"word\":\"meantime\",\"start\":0.62,\"end\":1.24}]"} {"utterance_id": "1188-133604-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0027.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.96, 1.8, "raphael"]], "response_mode": "offline", "answer": "[{\"word\":\"raphael\",\"start\":0.96,\"end\":1.8}]"} {"utterance_id": "5105-28240-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0008.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "You will hear a complete utterance. When it ends, list every instance of a weather term you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.26, 8.73, "weather"]], "response_mode": "offline", "answer": "[{\"word\":\"weather\",\"start\":8.26,\"end\":8.73}]"} {"utterance_id": "5683-32866-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0014.flac", "task_type": "lexical_detection", "query": "talking", "prompt": "After the audio ends, extract all occurrences of the word \"talking\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.91, 2.35, "talking"]], "response_mode": "offline", "answer": "[{\"word\":\"talking\",\"start\":1.91,\"end\":2.35}]"} {"utterance_id": "4446-2275-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0011.flac", "task_type": "lexical_detection", "query": "bent", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"bent\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.86, 1.13, "bent"]], "response_mode": "offline", "answer": "[{\"word\":\"bent\",\"start\":0.86,\"end\":1.13}]"} {"utterance_id": "121-127105-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0014.flac", "task_type": "lexical_detection", "query": "acute", "prompt": "After the audio ends, extract all occurrences of the word \"acute\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.31, 1.92, "acute"]], "response_mode": "offline", "answer": "[{\"word\":\"acute\",\"start\":1.31,\"end\":1.92}]"} {"utterance_id": "4970-29095-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0021.flac", "task_type": "lexical_counting", "query": "street", "prompt": "Process the full audio offline. For every occurrence of the word \"street\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.1, 3.6, "one"], [16.1, 16.42, "two"], [18.91, 19.38, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.1,\"end\":3.6},{\"word\":\"two\",\"start\":16.1,\"end\":16.42},{\"word\":\"three\",\"start\":18.91,\"end\":19.38}]"} {"utterance_id": "672-122797-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0007.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.14, 2.61, "cutters"]], "response_mode": "offline", "answer": "[{\"word\":\"cutters\",\"start\":2.14,\"end\":2.61}]"} {"utterance_id": "1580-141083-0049", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0049.flac", "task_type": "lexical_detection", "query": "anyone", "prompt": "When the utterance is complete, report all instances of the word \"anyone\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.44, 0.81, "anyone"]], "response_mode": "offline", "answer": "[{\"word\":\"anyone\",\"start\":0.44,\"end\":0.81}]"} {"utterance_id": "7729-102255-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0033.flac", "task_type": "lexical_detection", "query": "protect", "prompt": "After the audio ends, extract all occurrences of the word \"protect\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.51, 1.94, "protect"]], "response_mode": "offline", "answer": "[{\"word\":\"protect\",\"start\":1.51,\"end\":1.94}]"} {"utterance_id": "8224-274384-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0002.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.09, 9.57, "king"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":9.09,\"end\":9.57}]"} {"utterance_id": "3729-6852-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0026.flac", "task_type": "lexical_detection", "query": "exact", "prompt": "Once the audio has finished, identify every instance of the word \"exact\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.86, 4.51, "exact"]], "response_mode": "offline", "answer": "[{\"word\":\"exact\",\"start\":3.86,\"end\":4.51}]"} {"utterance_id": "61-70968-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0037.flac", "task_type": "lexical_detection", "query": "lording", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"lording\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.49, 2.08, "lording"]], "response_mode": "offline", "answer": "[{\"word\":\"lording\",\"start\":1.49,\"end\":2.08}]"} {"utterance_id": "7176-88083-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0025.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.13, 3.5, "head"]], "response_mode": "offline", "answer": "[{\"word\":\"head\",\"start\":3.13,\"end\":3.5}]"} {"utterance_id": "2094-142345-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0010.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen carefully to the entire clip, then report every instance of a family relation as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.72, 4.06, "aunt's"]], "response_mode": "offline", "answer": "[{\"word\":\"aunt's\",\"start\":3.72,\"end\":4.06}]"} {"utterance_id": "6829-68771-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0027.flac", "task_type": "lexical_detection", "query": "excuse", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"excuse\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.23, 4.7, "excuse"]], "response_mode": "offline", "answer": "[{\"word\":\"excuse\",\"start\":4.23,\"end\":4.7}]"} {"utterance_id": "2830-3980-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0024.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the full audio. Count every occurrence of a person's name. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.63, 1.92, "one"], [3.21, 3.69, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.63,\"end\":1.92},{\"word\":\"two\",\"start\":3.21,\"end\":3.69}]"} {"utterance_id": "7176-88083-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0025.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.57, 6.11, "fisherman"]], "response_mode": "offline", "answer": "[{\"word\":\"fisherman\",\"start\":5.57,\"end\":6.11}]"} {"utterance_id": "8455-210777-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0019.flac", "task_type": "lexical_detection", "query": "fathers", "prompt": "Once the audio has finished, identify every instance of the word \"fathers\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.39, 5.93, "fathers"]], "response_mode": "offline", "answer": "[{\"word\":\"fathers\",\"start\":5.39,\"end\":5.93}]"} {"utterance_id": "3729-6852-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0004.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.24, 10.76, "one"], [14.48, 14.72, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":10.24,\"end\":10.76},{\"word\":\"one\",\"start\":14.48,\"end\":14.72}]"} {"utterance_id": "7176-92135-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0037.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Process the full audio offline. Find every instance of a food name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.65, 1.93, "meal"]], "response_mode": "offline", "answer": "[{\"word\":\"meal\",\"start\":1.65,\"end\":1.93}]"} {"utterance_id": "4970-29095-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0032.flac", "task_type": "lexical_detection", "query": "talent", "prompt": "Process the full audio offline. Find every instance of the word \"talent\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.28, 2.65, "talent"]], "response_mode": "offline", "answer": "[{\"word\":\"talent\",\"start\":2.28,\"end\":2.65}]"} {"utterance_id": "260-123288-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0004.flac", "task_type": "lexical_detection", "query": "calm", "prompt": "When the utterance is complete, report all instances of the word \"calm\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.54, 4.14, "calm"]], "response_mode": "offline", "answer": "[{\"word\":\"calm\",\"start\":3.54,\"end\":4.14}]"} {"utterance_id": "121-121726-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0011.flac", "task_type": "lexical_detection", "query": "husband", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"husband\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 1.37, "husband"]], "response_mode": "offline", "answer": "[{\"word\":\"husband\",\"start\":0.51,\"end\":1.37}]"} {"utterance_id": "4446-2275-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0028.flac", "task_type": "lexical_detection", "query": "pulling", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"pulling\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.85, 2.13, "pulling"]], "response_mode": "offline", "answer": "[{\"word\":\"pulling\",\"start\":1.85,\"end\":2.13}]"} {"utterance_id": "8463-294825-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0009.flac", "task_type": "lexical_counting", "query": "yet", "prompt": "Process the full audio offline. For every occurrence of the word \"yet\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.78, 6.03, "one"], [12.61, 12.78, "two"], [17.14, 17.3, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.78,\"end\":6.03},{\"word\":\"two\",\"start\":12.61,\"end\":12.78},{\"word\":\"three\",\"start\":17.14,\"end\":17.3}]"} {"utterance_id": "8463-294828-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0006.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.69, 2.4, "scientists"]], "response_mode": "offline", "answer": "[{\"word\":\"scientists\",\"start\":1.69,\"end\":2.4}]"} {"utterance_id": "4992-23283-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0020.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an ordinal number with fields word, start, end (seconds). If none, reply [].", "labels": [[6.66, 7.16, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":6.66,\"end\":7.16}]"} {"utterance_id": "8555-292519-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0003.flac", "task_type": "lexical_detection", "query": "sails", "prompt": "After listening to the whole recording, output every instance of the word \"sails\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.88, 10.31, "sails"]], "response_mode": "offline", "answer": "[{\"word\":\"sails\",\"start\":9.88,\"end\":10.31}]"} {"utterance_id": "4446-2275-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0027.flac", "task_type": "lexical_detection", "query": "moved", "prompt": "After the audio ends, extract all occurrences of the word \"moved\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 0.85, "moved"]], "response_mode": "offline", "answer": "[{\"word\":\"moved\",\"start\":0.45,\"end\":0.85}]"} {"utterance_id": "260-123286-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0008.flac", "task_type": "lexical_detection", "query": "talk", "prompt": "Process the full audio offline. Find every instance of the word \"talk\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.03, 1.3, "talk"]], "response_mode": "offline", "answer": "[{\"word\":\"talk\",\"start\":1.03,\"end\":1.3}]"} {"utterance_id": "5142-36377-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0016.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.07, 2.66, "prophet"]], "response_mode": "offline", "answer": "[{\"word\":\"prophet\",\"start\":2.07,\"end\":2.66}]"} {"utterance_id": "1320-122617-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0035.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.09, 3.69, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":3.09,\"end\":3.69}]"} {"utterance_id": "672-122797-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0005.flac", "task_type": "lexical_detection", "query": "angry", "prompt": "Process the full audio offline. Find every instance of the word \"angry\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.23, 2.75, "angry"]], "response_mode": "offline", "answer": "[{\"word\":\"angry\",\"start\":2.23,\"end\":2.75}]"} {"utterance_id": "61-70970-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0014.flac", "task_type": "lexical_detection", "query": "crossed", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"crossed\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.83, 1.16, "crossed"]], "response_mode": "offline", "answer": "[{\"word\":\"crossed\",\"start\":0.83,\"end\":1.16}]"} {"utterance_id": "1284-1180-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0031.flac", "task_type": "semantic_detection", "query": "city", "prompt": "After listening to the whole recording, output every instance of a city name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 1.29, "emerald city"]], "response_mode": "offline", "answer": "[{\"word\":\"emerald city\",\"start\":0.53,\"end\":1.29}]"} {"utterance_id": "5142-33396-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0050.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.86, 1.23, "heart's"]], "response_mode": "offline", "answer": "[{\"word\":\"heart's\",\"start\":0.86,\"end\":1.23}]"} {"utterance_id": "2961-961-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0005.flac", "task_type": "lexical_detection", "query": "solon", "prompt": "Process the full audio offline. Find every instance of the word \"solon\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.28, 1.82, "solon"]], "response_mode": "offline", "answer": "[{\"word\":\"solon\",\"start\":1.28,\"end\":1.82}]"} {"utterance_id": "5683-32866-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0026.flac", "task_type": "lexical_detection", "query": "boldly", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"boldly\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.5, 1.02, "boldly"]], "response_mode": "offline", "answer": "[{\"word\":\"boldly\",\"start\":0.5,\"end\":1.02}]"} {"utterance_id": "1089-134691-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0001.flac", "task_type": "lexical_detection", "query": "paced", "prompt": "When the utterance is complete, report all instances of the word \"paced\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.52, 1.96, "paced"]], "response_mode": "offline", "answer": "[{\"word\":\"paced\",\"start\":1.52,\"end\":1.96}]"} {"utterance_id": "4077-13754-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0014.flac", "task_type": "lexical_detection", "query": "measures", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"measures\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.51, 1.95, "measures"]], "response_mode": "offline", "answer": "[{\"word\":\"measures\",\"start\":1.51,\"end\":1.95}]"} {"utterance_id": "6930-81414-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0014.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.43, 4.88, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":4.43,\"end\":4.88}]"} {"utterance_id": "4077-13751-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0018.flac", "task_type": "lexical_detection", "query": "horror", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"horror\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.82, 3.33, "horror"]], "response_mode": "offline", "answer": "[{\"word\":\"horror\",\"start\":2.82,\"end\":3.33}]"} {"utterance_id": "2961-960-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0019.flac", "task_type": "lexical_detection", "query": "poet", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"poet\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[7.88, 8.4, "poet"]], "response_mode": "offline", "answer": "[{\"word\":\"poet\",\"start\":7.88,\"end\":8.4}]"} {"utterance_id": "4992-23283-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0018.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Process the full audio offline. For every occurrence of a body part, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.17, 2.51, "one"], [3.62, 3.93, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.17,\"end\":2.51},{\"word\":\"two\",\"start\":3.62,\"end\":3.93}]"} {"utterance_id": "1320-122612-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0009.flac", "task_type": "lexical_detection", "query": "wonderful", "prompt": "When the utterance is complete, report all instances of the word \"wonderful\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.47, 1.97, "wonderful"]], "response_mode": "offline", "answer": "[{\"word\":\"wonderful\",\"start\":1.47,\"end\":1.97}]"} {"utterance_id": "2300-131720-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0017.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "When the utterance is complete, report all instances of an ordinal number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.54, 2.95, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":2.54,\"end\":2.95}]"} {"utterance_id": "4970-29093-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0018.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.13, 6.51, "henry's"]], "response_mode": "offline", "answer": "[{\"word\":\"henry's\",\"start\":6.13,\"end\":6.51}]"} {"utterance_id": "3570-5694-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0013.flac", "task_type": "lexical_detection", "query": "inheritance", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"inheritance\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.05, 2.63, "inheritance"], [3.96, 4.49, "inheritance"]], "response_mode": "offline", "answer": "[{\"word\":\"inheritance\",\"start\":2.05,\"end\":2.63},{\"word\":\"inheritance\",\"start\":3.96,\"end\":4.49}]"} {"utterance_id": "1089-134686-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0025.flac", "task_type": "lexical_detection", "query": "gentle", "prompt": "Process the full audio offline. Find every instance of the word \"gentle\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.62, 0.94, "gentle"]], "response_mode": "offline", "answer": "[{\"word\":\"gentle\",\"start\":0.62,\"end\":0.94}]"} {"utterance_id": "672-122797-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0037.flac", "task_type": "lexical_detection", "query": "that's", "prompt": "Once the audio has finished, identify every instance of the word \"that's\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.1, 0.53, "that's"]], "response_mode": "offline", "answer": "[{\"word\":\"that's\",\"start\":0.1,\"end\":0.53}]"} {"utterance_id": "4507-16021-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0031.flac", "task_type": "lexical_detection", "query": "choice", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"choice\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.42, 12.93, "choice"]], "response_mode": "offline", "answer": "[{\"word\":\"choice\",\"start\":12.42,\"end\":12.93}]"} {"utterance_id": "8463-294828-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0026.flac", "task_type": "lexical_detection", "query": "commander", "prompt": "When the utterance is complete, report all instances of the word \"commander\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.56, 1.14, "commander"]], "response_mode": "offline", "answer": "[{\"word\":\"commander\",\"start\":0.56,\"end\":1.14}]"} {"utterance_id": "2300-131720-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0000.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Once the audio has finished, identify every instance of a plant name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.73, 1.14, "plant"]], "response_mode": "offline", "answer": "[{\"word\":\"plant\",\"start\":0.73,\"end\":1.14}]"} {"utterance_id": "5683-32866-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0017.flac", "task_type": "lexical_detection", "query": "uneasy", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"uneasy\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.81, 1.35, "uneasy"]], "response_mode": "offline", "answer": "[{\"word\":\"uneasy\",\"start\":0.81,\"end\":1.35}]"} {"utterance_id": "1284-1180-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0017.flac", "task_type": "lexical_detection", "query": "fireplace", "prompt": "Once the audio has finished, identify every instance of the word \"fireplace\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.31, 2.02, "fireplace"]], "response_mode": "offline", "answer": "[{\"word\":\"fireplace\",\"start\":1.31,\"end\":2.02}]"} {"utterance_id": "7176-88083-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0022.flac", "task_type": "lexical_detection", "query": "loss", "prompt": "After the audio ends, extract all occurrences of the word \"loss\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.7, 2.04, "loss"]], "response_mode": "offline", "answer": "[{\"word\":\"loss\",\"start\":1.7,\"end\":2.04}]"} {"utterance_id": "260-123440-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0019.flac", "task_type": "lexical_detection", "query": "mouse", "prompt": "Once the audio has finished, identify every instance of the word \"mouse\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.79, 2.07, "mouse"]], "response_mode": "offline", "answer": "[{\"word\":\"mouse\",\"start\":1.79,\"end\":2.07}]"} {"utterance_id": "5105-28233-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28233/5105-28233-0002.flac", "task_type": "lexical_detection", "query": "literary", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"literary\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.31, 5.94, "literary"]], "response_mode": "offline", "answer": "[{\"word\":\"literary\",\"start\":5.31,\"end\":5.94}]"} {"utterance_id": "4970-29095-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0036.flac", "task_type": "lexical_counting", "query": "career", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"career\" with word/start/end. If none, reply [].", "labels": [[2.04, 2.53, "one"], [3.67, 4.06, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.04,\"end\":2.53},{\"word\":\"two\",\"start\":3.67,\"end\":4.06}]"} {"utterance_id": "7127-75947-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0040.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "Process the full audio offline. For every occurrence of a plant name, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.21, 8.64, "one"], [9.83, 10.39, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":8.21,\"end\":8.64},{\"word\":\"two\",\"start\":9.83,\"end\":10.39}]"} {"utterance_id": "5142-33396-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0006.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "When the utterance is complete, report how a number accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.87, 1.28, "one"], [4.9, 5.35, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.87,\"end\":1.28},{\"word\":\"two\",\"start\":4.9,\"end\":5.35}]"} {"utterance_id": "4077-13751-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0006.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After listening to the whole recording, output every instance of an organization name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.89, 11.28, "church"]], "response_mode": "offline", "answer": "[{\"word\":\"church\",\"start\":10.89,\"end\":11.28}]"} {"utterance_id": "121-127105-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0020.flac", "task_type": "lexical_detection", "query": "literal", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"literal\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.64, 12.18, "literal"]], "response_mode": "offline", "answer": "[{\"word\":\"literal\",\"start\":11.64,\"end\":12.18}]"} {"utterance_id": "1995-1826-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0023.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "When the utterance is complete, report how a plant name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.15, 2.7, "one"], [4.74, 5.0, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.15,\"end\":2.7},{\"word\":\"two\",\"start\":4.74,\"end\":5.0}]"} {"utterance_id": "3570-5696-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0006.flac", "task_type": "lexical_detection", "query": "everyday", "prompt": "After listening to the whole recording, output every instance of the word \"everyday\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.23, 1.61, "everyday"]], "response_mode": "offline", "answer": "[{\"word\":\"everyday\",\"start\":1.23,\"end\":1.61}]"} {"utterance_id": "7729-102255-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0006.flac", "task_type": "semantic_counting", "query": "city", "prompt": "Listen to the entire clip, then report the count sequence for a city name as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.06, 2.68, "one"], [10.6, 11.23, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.06,\"end\":2.68},{\"word\":\"two\",\"start\":10.6,\"end\":11.23}]"} {"utterance_id": "5683-32879-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0003.flac", "task_type": "lexical_detection", "query": "hide", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"hide\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.98, 1.33, "hide"]], "response_mode": "offline", "answer": "[{\"word\":\"hide\",\"start\":0.98,\"end\":1.33}]"} {"utterance_id": "5639-40744-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0037.flac", "task_type": "lexical_counting", "query": "real", "prompt": "Once the audio has finished, list each occurrence of the word \"real\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[3.87, 4.16, "one"], [5.22, 5.57, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.87,\"end\":4.16},{\"word\":\"two\",\"start\":5.22,\"end\":5.57}]"} {"utterance_id": "8224-274384-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0009.flac", "task_type": "lexical_detection", "query": "parliament", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"parliament\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.6, 1.3, "parliament"]], "response_mode": "offline", "answer": "[{\"word\":\"parliament\",\"start\":0.6,\"end\":1.3}]"} {"utterance_id": "5142-36586-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36586/5142-36586-0001.flac", "task_type": "lexical_detection", "query": "lower", "prompt": "After the audio ends, extract all occurrences of the word \"lower\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.12, 1.41, "lower"]], "response_mode": "offline", "answer": "[{\"word\":\"lower\",\"start\":1.12,\"end\":1.41}]"} {"utterance_id": "7729-102255-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0000.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.91, 2.54, "thirty six"]], "response_mode": "offline", "answer": "[{\"word\":\"thirty six\",\"start\":1.91,\"end\":2.54}]"} {"utterance_id": "3570-5694-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0012.flac", "task_type": "lexical_detection", "query": "elaborate", "prompt": "Once the audio has finished, identify every instance of the word \"elaborate\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.04, 1.57, "elaborate"]], "response_mode": "offline", "answer": "[{\"word\":\"elaborate\",\"start\":1.04,\"end\":1.57}]"} {"utterance_id": "260-123288-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0024.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "When the utterance is complete, report all instances of a weather term including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.46, 14.17, "whirlwind"]], "response_mode": "offline", "answer": "[{\"word\":\"whirlwind\",\"start\":13.46,\"end\":14.17}]"} {"utterance_id": "3729-6852-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0004.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Once the audio has finished, list each occurrence of a number as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[10.24, 10.76, "one"], [14.48, 14.72, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":10.24,\"end\":10.76},{\"word\":\"two\",\"start\":14.48,\"end\":14.72}]"} {"utterance_id": "2830-3980-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0024.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.63, 1.92, "paul"], [3.21, 3.69, "christ"]], "response_mode": "offline", "answer": "[{\"word\":\"paul\",\"start\":1.63,\"end\":1.92},{\"word\":\"christ\",\"start\":3.21,\"end\":3.69}]"} {"utterance_id": "1089-134691-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0021.flac", "task_type": "lexical_detection", "query": "rude", "prompt": "Once the audio has finished, identify every instance of the word \"rude\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.67, 2.99, "rude"]], "response_mode": "offline", "answer": "[{\"word\":\"rude\",\"start\":2.67,\"end\":2.99}]"} {"utterance_id": "3575-170457-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0018.flac", "task_type": "semantic_detection", "query": "date", "prompt": "After the audio ends, extract all occurrences of a date. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.52, 2.49, "march sixteenth"]], "response_mode": "offline", "answer": "[{\"word\":\"march sixteenth\",\"start\":1.52,\"end\":2.49}]"} {"utterance_id": "7021-85628-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0022.flac", "task_type": "lexical_detection", "query": "mouths", "prompt": "Process the full audio offline. Find every instance of the word \"mouths\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.05, 4.51, "mouths"]], "response_mode": "offline", "answer": "[{\"word\":\"mouths\",\"start\":4.05,\"end\":4.51}]"} {"utterance_id": "5105-28233-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28233/5105-28233-0006.flac", "task_type": "lexical_detection", "query": "itself", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"itself\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.33, 2.85, "itself"]], "response_mode": "offline", "answer": "[{\"word\":\"itself\",\"start\":2.33,\"end\":2.85}]"} {"utterance_id": "7021-79730-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0006.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.68, 1.14, "mother"], [6.95, 7.5, "child"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":0.68,\"end\":1.14},{\"word\":\"child\",\"start\":6.95,\"end\":7.5}]"} {"utterance_id": "1188-133604-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0007.flac", "task_type": "lexical_detection", "query": "outer", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"outer\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.38, 8.61, "outer"]], "response_mode": "offline", "answer": "[{\"word\":\"outer\",\"start\":8.38,\"end\":8.61}]"} {"utterance_id": "8555-284447-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0015.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen carefully to the entire clip, then report every instance of an animal name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.64, 9.97, "goat"]], "response_mode": "offline", "answer": "[{\"word\":\"goat\",\"start\":9.64,\"end\":9.97}]"} {"utterance_id": "2830-3980-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0007.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "Listen to the entire clip, then report the count sequence for a profession as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.74, 1.27, "one"], [5.52, 5.94, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.74,\"end\":1.27},{\"word\":\"two\",\"start\":5.52,\"end\":5.94}]"} {"utterance_id": "1089-134686-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0009.flac", "task_type": "lexical_detection", "query": "beggar", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"beggar\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.79, 3.34, "beggar"]], "response_mode": "offline", "answer": "[{\"word\":\"beggar\",\"start\":2.79,\"end\":3.34}]"} {"utterance_id": "8555-292519-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0007.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Once the audio has finished, identify every instance of a weather term. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.73, 4.25, "clouds"]], "response_mode": "offline", "answer": "[{\"word\":\"clouds\",\"start\":3.73,\"end\":4.25}]"} {"utterance_id": "672-122797-0064", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0064.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for an animal name with word/start/end. If none, reply [].", "labels": [[1.79, 2.31, "one"], [11.32, 11.66, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.79,\"end\":2.31},{\"word\":\"two\",\"start\":11.32,\"end\":11.66}]"} {"utterance_id": "6829-68769-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0040.flac", "task_type": "lexical_detection", "query": "interview", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"interview\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.91, 2.32, "interview"]], "response_mode": "offline", "answer": "[{\"word\":\"interview\",\"start\":1.91,\"end\":2.32}]"} {"utterance_id": "1580-141084-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0030.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.45, 1.94, "holmes"]], "response_mode": "offline", "answer": "[{\"word\":\"holmes\",\"start\":1.45,\"end\":1.94}]"} {"utterance_id": "6829-68769-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0018.flac", "task_type": "lexical_detection", "query": "passage", "prompt": "Process the full audio offline. Find every instance of the word \"passage\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.71, 6.19, "passage"]], "response_mode": "offline", "answer": "[{\"word\":\"passage\",\"start\":5.71,\"end\":6.19}]"} {"utterance_id": "2830-3980-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0040.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.16, 1.63, "apostle"]], "response_mode": "offline", "answer": "[{\"word\":\"apostle\",\"start\":1.16,\"end\":1.63}]"} {"utterance_id": "237-126133-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.4, 2.93, "mister king"]], "response_mode": "offline", "answer": "[{\"word\":\"mister king\",\"start\":2.4,\"end\":2.93}]"} {"utterance_id": "1995-1836-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0011.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.0, 4.58, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":4.0,\"end\":4.58}]"} {"utterance_id": "1995-1836-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0007.flac", "task_type": "lexical_detection", "query": "education", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"education\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.37, 2.06, "education"]], "response_mode": "offline", "answer": "[{\"word\":\"education\",\"start\":1.37,\"end\":2.06}]"} {"utterance_id": "5105-28240-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0009.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.77, 8.96, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":8.77,\"end\":8.96}]"} {"utterance_id": "2830-3980-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0028.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.66, 3.28, "apostles"]], "response_mode": "offline", "answer": "[{\"word\":\"apostles\",\"start\":2.66,\"end\":3.28}]"} {"utterance_id": "4992-41806-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0016.flac", "task_type": "lexical_detection", "query": "piano", "prompt": "Process the full audio offline. Find every instance of the word \"piano\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.62, 8.22, "piano"]], "response_mode": "offline", "answer": "[{\"word\":\"piano\",\"start\":7.62,\"end\":8.22}]"} {"utterance_id": "8555-292519-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0001.flac", "task_type": "lexical_detection", "query": "guided", "prompt": "After the audio ends, extract all occurrences of the word \"guided\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 1.03, "guided"]], "response_mode": "offline", "answer": "[{\"word\":\"guided\",\"start\":0.53,\"end\":1.03}]"} {"utterance_id": "2094-142345-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0030.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an animal name with fields word, start, end (seconds). If none, reply [].", "labels": [[15.21, 15.58, "pig"]], "response_mode": "offline", "answer": "[{\"word\":\"pig\",\"start\":15.21,\"end\":15.58}]"} {"utterance_id": "2961-961-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0002.flac", "task_type": "lexical_detection", "query": "citizen", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"citizen\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.38, 3.84, "citizen"]], "response_mode": "offline", "answer": "[{\"word\":\"citizen\",\"start\":3.38,\"end\":3.84}]"} {"utterance_id": "6930-81414-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0006.flac", "task_type": "lexical_detection", "query": "distinctly", "prompt": "Process the full audio offline. Find every instance of the word \"distinctly\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.28, 5.11, "distinctly"]], "response_mode": "offline", "answer": "[{\"word\":\"distinctly\",\"start\":4.28,\"end\":5.11}]"} {"utterance_id": "908-157963-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0018.flac", "task_type": "lexical_detection", "query": "thou", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"thou\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.97, 1.18, "thou"]], "response_mode": "offline", "answer": "[{\"word\":\"thou\",\"start\":0.97,\"end\":1.18}]"} {"utterance_id": "1580-141084-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0020.flac", "task_type": "lexical_detection", "query": "held", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"held\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.74, 1.98, "held"]], "response_mode": "offline", "answer": "[{\"word\":\"held\",\"start\":1.74,\"end\":1.98}]"} {"utterance_id": "672-122797-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0015.flac", "task_type": "lexical_detection", "query": "warm", "prompt": "After listening to the whole recording, output every instance of the word \"warm\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.08, 1.58, "warm"]], "response_mode": "offline", "answer": "[{\"word\":\"warm\",\"start\":1.08,\"end\":1.58}]"} {"utterance_id": "3575-170457-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0048.flac", "task_type": "lexical_detection", "query": "advice", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"advice\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.36, 3.0, "advice"]], "response_mode": "offline", "answer": "[{\"word\":\"advice\",\"start\":2.36,\"end\":3.0}]"} {"utterance_id": "2830-3979-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[8.26, 8.54, "luke"]], "response_mode": "offline", "answer": "[{\"word\":\"luke\",\"start\":8.26,\"end\":8.54}]"} {"utterance_id": "121-127105-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0026.flac", "task_type": "lexical_detection", "query": "statement", "prompt": "Process the full audio offline. Find every instance of the word \"statement\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.55, 3.18, "statement"]], "response_mode": "offline", "answer": "[{\"word\":\"statement\",\"start\":2.55,\"end\":3.18}]"} {"utterance_id": "1580-141084-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0050.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.44, 1.22, "mister soames"]], "response_mode": "offline", "answer": "[{\"word\":\"mister soames\",\"start\":0.44,\"end\":1.22}]"} {"utterance_id": "3729-6852-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0046.flac", "task_type": "lexical_detection", "query": "housekeeper", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"housekeeper\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.56, 1.2, "housekeeper"]], "response_mode": "offline", "answer": "[{\"word\":\"housekeeper\",\"start\":0.56,\"end\":1.2}]"} {"utterance_id": "4077-13754-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0008.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Do not answer until the audio ends. Then list each instance of a country name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.46, 3.99, "lands"], [7.0, 7.4, "nations"]], "response_mode": "offline", "answer": "[{\"word\":\"lands\",\"start\":3.46,\"end\":3.99},{\"word\":\"nations\",\"start\":7.0,\"end\":7.4}]"} {"utterance_id": "8230-279154-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0012.flac", "task_type": "lexical_detection", "query": "familiarity", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"familiarity\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.28, 1.25, "familiarity"]], "response_mode": "offline", "answer": "[{\"word\":\"familiarity\",\"start\":0.28,\"end\":1.25}]"} {"utterance_id": "4446-2273-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0031.flac", "task_type": "lexical_detection", "query": "front", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"front\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.26, 1.65, "front"]], "response_mode": "offline", "answer": "[{\"word\":\"front\",\"start\":1.26,\"end\":1.65}]"} {"utterance_id": "6829-68769-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0026.flac", "task_type": "lexical_detection", "query": "narrow", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"narrow\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.95, 3.28, "narrow"]], "response_mode": "offline", "answer": "[{\"word\":\"narrow\",\"start\":2.95,\"end\":3.28}]"} {"utterance_id": "5142-36377-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0023.flac", "task_type": "lexical_detection", "query": "john", "prompt": "Process the full audio offline. Find every instance of the word \"john\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.6, 3.85, "john"]], "response_mode": "offline", "answer": "[{\"word\":\"john\",\"start\":3.6,\"end\":3.85}]"} {"utterance_id": "908-31957-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0022.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "Process the full audio offline. Find every instance of a disease name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.05, 3.51, "ills"]], "response_mode": "offline", "answer": "[{\"word\":\"ills\",\"start\":3.05,\"end\":3.51}]"} {"utterance_id": "8230-279154-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0038.flac", "task_type": "semantic_counting", "query": "ordinal", "prompt": "Once the audio has finished, list each occurrence of an ordinal number as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[11.65, 12.13, "one"], [20.28, 20.76, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":11.65,\"end\":12.13},{\"word\":\"two\",\"start\":20.28,\"end\":20.76}]"} {"utterance_id": "8455-210777-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0017.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen carefully to the entire clip, then report every instance of a food name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.47, 4.85, "dinner"]], "response_mode": "offline", "answer": "[{\"word\":\"dinner\",\"start\":4.47,\"end\":4.85}]"} {"utterance_id": "2094-142345-0060", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0060.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Once the audio has finished, identify every instance of a money or currency term. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.68, 2.11, "capital"]], "response_mode": "offline", "answer": "[{\"word\":\"capital\",\"start\":1.68,\"end\":2.11}]"} {"utterance_id": "1580-141084-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0025.flac", "task_type": "lexical_detection", "query": "think", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"think\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.51, 1.74, "think"]], "response_mode": "offline", "answer": "[{\"word\":\"think\",\"start\":1.51,\"end\":1.74}]"} {"utterance_id": "237-134493-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0012.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.09, 1.49, "knees"]], "response_mode": "offline", "answer": "[{\"word\":\"knees\",\"start\":1.09,\"end\":1.49}]"} {"utterance_id": "672-122797-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0033.flac", "task_type": "lexical_detection", "query": "story", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"story\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.57, 1.19, "story"]], "response_mode": "offline", "answer": "[{\"word\":\"story\",\"start\":0.57,\"end\":1.19}]"} {"utterance_id": "1320-122617-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0008.flac", "task_type": "lexical_detection", "query": "fear", "prompt": "After the audio ends, extract all occurrences of the word \"fear\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.62, 2.87, "fear"]], "response_mode": "offline", "answer": "[{\"word\":\"fear\",\"start\":2.62,\"end\":2.87}]"} {"utterance_id": "2830-3980-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0007.flac", "task_type": "lexical_detection", "query": "minister", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"minister\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.52, 5.94, "minister"]], "response_mode": "offline", "answer": "[{\"word\":\"minister\",\"start\":5.52,\"end\":5.94}]"} {"utterance_id": "908-157963-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0024.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "After listening to the whole recording, output every instance of an animal name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.5, 3.11, "worm"]], "response_mode": "offline", "answer": "[{\"word\":\"worm\",\"start\":2.5,\"end\":3.11}]"} {"utterance_id": "7176-88083-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0007.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen carefully to the entire clip, then report every instance of a color name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.08, 9.4, "red"]], "response_mode": "offline", "answer": "[{\"word\":\"red\",\"start\":9.08,\"end\":9.4}]"} {"utterance_id": "1221-135767-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0009.flac", "task_type": "lexical_detection", "query": "frowning", "prompt": "When the utterance is complete, report all instances of the word \"frowning\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.28, 3.88, "frowning"]], "response_mode": "offline", "answer": "[{\"word\":\"frowning\",\"start\":3.28,\"end\":3.88}]"} {"utterance_id": "1284-1180-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0026.flac", "task_type": "semantic_detection", "query": "food", "prompt": "You will hear a complete utterance. When it ends, list every instance of a food name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.62, 7.05, "meals"]], "response_mode": "offline", "answer": "[{\"word\":\"meals\",\"start\":6.62,\"end\":7.05}]"} {"utterance_id": "2094-142345-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0033.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.03, 7.55, "rachel"]], "response_mode": "offline", "answer": "[{\"word\":\"rachel\",\"start\":7.03,\"end\":7.55}]"} {"utterance_id": "1188-133604-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0034.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Do not answer until the audio ends. Then list each instance of an organization name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.97, 6.38, "schools"]], "response_mode": "offline", "answer": "[{\"word\":\"schools\",\"start\":5.97,\"end\":6.38}]"} {"utterance_id": "7176-88083-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0021.flac", "task_type": "lexical_detection", "query": "strike", "prompt": "Once the audio has finished, identify every instance of the word \"strike\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[11.6, 12.2, "strike"]], "response_mode": "offline", "answer": "[{\"word\":\"strike\",\"start\":11.6,\"end\":12.2}]"} {"utterance_id": "7021-79730-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0003.flac", "task_type": "lexical_detection", "query": "chaise", "prompt": "After the audio ends, extract all occurrences of the word \"chaise\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.61, 0.99, "chaise"]], "response_mode": "offline", "answer": "[{\"word\":\"chaise\",\"start\":0.61,\"end\":0.99}]"} {"utterance_id": "1089-134691-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0016.flac", "task_type": "lexical_detection", "query": "voyaging", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"voyaging\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.79, 1.31, "voyaging"], [5.54, 6.07, "voyaging"]], "response_mode": "offline", "answer": "[{\"word\":\"voyaging\",\"start\":0.79,\"end\":1.31},{\"word\":\"voyaging\",\"start\":5.54,\"end\":6.07}]"} {"utterance_id": "1284-1180-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0008.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a time of day with fields word, start, end (seconds). If none, reply [].", "labels": [[0.73, 1.14, "morning"], [3.15, 3.57, "noon"]], "response_mode": "offline", "answer": "[{\"word\":\"morning\",\"start\":0.73,\"end\":1.14},{\"word\":\"noon\",\"start\":3.15,\"end\":3.57}]"} {"utterance_id": "237-126133-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0025.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.65, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":2.35,\"end\":2.65}]"} {"utterance_id": "1580-141083-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0035.flac", "task_type": "lexical_detection", "query": "what's", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"what's\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.11, 4.37, "what's"]], "response_mode": "offline", "answer": "[{\"word\":\"what's\",\"start\":4.11,\"end\":4.37}]"} {"utterance_id": "8230-279154-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0009.flac", "task_type": "lexical_counting", "query": "image", "prompt": "Listen to the full audio. Count every occurrence of the word \"image\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.91, 5.52, "one"], [9.98, 10.34, "two"], [16.12, 16.82, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.91,\"end\":5.52},{\"word\":\"two\",\"start\":9.98,\"end\":10.34},{\"word\":\"three\",\"start\":16.12,\"end\":16.82}]"} {"utterance_id": "237-134500-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0020.flac", "task_type": "lexical_detection", "query": "yes", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"yes\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.51, 1.01, "yes"]], "response_mode": "offline", "answer": "[{\"word\":\"yes\",\"start\":0.51,\"end\":1.01}]"} {"utterance_id": "4446-2275-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0038.flac", "task_type": "lexical_detection", "query": "ask", "prompt": "Process the full audio offline. Find every instance of the word \"ask\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.55, 0.86, "ask"]], "response_mode": "offline", "answer": "[{\"word\":\"ask\",\"start\":0.55,\"end\":0.86}]"} {"utterance_id": "8455-210777-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0052.flac", "task_type": "lexical_detection", "query": "allow", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"allow\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.85, 1.12, "allow"]], "response_mode": "offline", "answer": "[{\"word\":\"allow\",\"start\":0.85,\"end\":1.12}]"} {"utterance_id": "672-122797-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0018.flac", "task_type": "lexical_detection", "query": "sunlight", "prompt": "When the utterance is complete, report all instances of the word \"sunlight\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.11, 4.74, "sunlight"]], "response_mode": "offline", "answer": "[{\"word\":\"sunlight\",\"start\":4.11,\"end\":4.74}]"} {"utterance_id": "2961-960-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0014.flac", "task_type": "lexical_detection", "query": "fishes", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"fishes\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.95, 8.61, "fishes"]], "response_mode": "offline", "answer": "[{\"word\":\"fishes\",\"start\":7.95,\"end\":8.61}]"} {"utterance_id": "1284-1180-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0026.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.91, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":0.53,\"end\":0.91}]"} {"utterance_id": "260-123286-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0015.flac", "task_type": "lexical_detection", "query": "atlantic", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"atlantic\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.88, 3.52, "atlantic"]], "response_mode": "offline", "answer": "[{\"word\":\"atlantic\",\"start\":2.88,\"end\":3.52}]"} {"utterance_id": "260-123288-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0010.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Once the audio has finished, identify every instance of a weather term. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.46, 4.62, "saint elmo's fire"], [7.51, 7.95, "wind"]], "response_mode": "offline", "answer": "[{\"word\":\"saint elmo's fire\",\"start\":3.46,\"end\":4.62},{\"word\":\"wind\",\"start\":7.51,\"end\":7.95}]"} {"utterance_id": "1320-122617-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0021.flac", "task_type": "lexical_detection", "query": "singer", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"singer\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.59, 3.95, "singer"]], "response_mode": "offline", "answer": "[{\"word\":\"singer\",\"start\":3.59,\"end\":3.95}]"} {"utterance_id": "8455-210777-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0048.flac", "task_type": "lexical_detection", "query": "kidnap", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"kidnap\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.18, 2.64, "kidnap"]], "response_mode": "offline", "answer": "[{\"word\":\"kidnap\",\"start\":2.18,\"end\":2.64}]"} {"utterance_id": "8230-279154-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0033.flac", "task_type": "lexical_counting", "query": "time", "prompt": "When the utterance is complete, report how the word \"time\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.84, 5.2, "one"], [10.9, 11.27, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.84,\"end\":5.2},{\"word\":\"two\",\"start\":10.9,\"end\":11.27}]"} {"utterance_id": "1580-141084-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0020.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.29, 2.74, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":2.29,\"end\":2.74}]"} {"utterance_id": "1284-1181-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0011.flac", "task_type": "semantic_detection", "query": "color", "prompt": "You will hear a complete utterance. When it ends, list every instance of a color name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.6, 5.06, "golden"], [6.59, 6.97, "golden"]], "response_mode": "offline", "answer": "[{\"word\":\"golden\",\"start\":4.6,\"end\":5.06},{\"word\":\"golden\",\"start\":6.59,\"end\":6.97}]"} {"utterance_id": "2830-3980-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0046.flac", "task_type": "lexical_detection", "query": "enough", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"enough\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.82, 1.11, "enough"]], "response_mode": "offline", "answer": "[{\"word\":\"enough\",\"start\":0.82,\"end\":1.11}]"} {"utterance_id": "6829-68771-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0019.flac", "task_type": "lexical_detection", "query": "costume", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"costume\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.18, 2.67, "costume"]], "response_mode": "offline", "answer": "[{\"word\":\"costume\",\"start\":2.18,\"end\":2.67}]"} {"utterance_id": "3570-5695-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0001.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "When the utterance is complete, report all instances of an ordinal number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.39, 1.75, "first"], [13.91, 14.42, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":1.39,\"end\":1.75},{\"word\":\"last\",\"start\":13.91,\"end\":14.42}]"} {"utterance_id": "8224-274381-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.86, 12.9, "lord elcho"]], "response_mode": "offline", "answer": "[{\"word\":\"lord elcho\",\"start\":11.86,\"end\":12.9}]"} {"utterance_id": "2961-961-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0019.flac", "task_type": "semantic_detection", "query": "country", "prompt": "After the audio ends, extract all occurrences of a country name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.22, 10.72, "nations"]], "response_mode": "offline", "answer": "[{\"word\":\"nations\",\"start\":10.22,\"end\":10.72}]"} {"utterance_id": "121-123852-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123852/121-123852-0003.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[15.72, 16.23, "nought"]], "response_mode": "offline", "answer": "[{\"word\":\"nought\",\"start\":15.72,\"end\":16.23}]"} {"utterance_id": "61-70970-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0007.flac", "task_type": "lexical_detection", "query": "clerk", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"clerk\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.87, 2.25, "clerk"]], "response_mode": "offline", "answer": "[{\"word\":\"clerk\",\"start\":1.87,\"end\":2.25}]"} {"utterance_id": "8555-284447-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0020.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "When the utterance is complete, report all instances of an animal name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.65, 0.98, "goat's"]], "response_mode": "offline", "answer": "[{\"word\":\"goat's\",\"start\":0.65,\"end\":0.98}]"} {"utterance_id": "1580-141083-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0030.flac", "task_type": "lexical_detection", "query": "flood", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"flood\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.35, 2.55, "flood"]], "response_mode": "offline", "answer": "[{\"word\":\"flood\",\"start\":2.35,\"end\":2.55}]"} {"utterance_id": "2830-3979-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0011.flac", "task_type": "lexical_counting", "query": "power", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"power\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[1.62, 1.98, "one"], [5.39, 5.76, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.62,\"end\":1.98},{\"word\":\"two\",\"start\":5.39,\"end\":5.76}]"} {"utterance_id": "5105-28240-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0012.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.3, 2.74, "captain"]], "response_mode": "offline", "answer": "[{\"word\":\"captain\",\"start\":2.3,\"end\":2.74}]"} {"utterance_id": "1995-1826-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0017.flac", "task_type": "lexical_detection", "query": "bit", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"bit\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.85, 1.01, "bit"]], "response_mode": "offline", "answer": "[{\"word\":\"bit\",\"start\":0.85,\"end\":1.01}]"} {"utterance_id": "7176-88083-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0023.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen to the full audio. After it finishes, find every instance of an animal name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.3, 8.61, "fish"]], "response_mode": "offline", "answer": "[{\"word\":\"fish\",\"start\":8.3,\"end\":8.61}]"} {"utterance_id": "7127-75947-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0040.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.21, 8.64, "brake"], [9.83, 10.39, "branch"]], "response_mode": "offline", "answer": "[{\"word\":\"brake\",\"start\":8.21,\"end\":8.64},{\"word\":\"branch\",\"start\":9.83,\"end\":10.39}]"} {"utterance_id": "5683-32879-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0007.flac", "task_type": "lexical_detection", "query": "struck", "prompt": "Process the full audio offline. Find every instance of the word \"struck\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.1, 4.53, "struck"]], "response_mode": "offline", "answer": "[{\"word\":\"struck\",\"start\":4.1,\"end\":4.53}]"} {"utterance_id": "7176-88083-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0025.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.34, 2.71, "bird"]], "response_mode": "offline", "answer": "[{\"word\":\"bird\",\"start\":2.34,\"end\":2.71}]"} {"utterance_id": "5142-33396-0066", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0066.flac", "task_type": "lexical_detection", "query": "frowning", "prompt": "Process the full audio offline. Find every instance of the word \"frowning\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.52, 2.11, "frowning"]], "response_mode": "offline", "answer": "[{\"word\":\"frowning\",\"start\":1.52,\"end\":2.11}]"} {"utterance_id": "1995-1826-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0000.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.09, 3.53, "fifteenth"]], "response_mode": "offline", "answer": "[{\"word\":\"fifteenth\",\"start\":3.09,\"end\":3.53}]"} {"utterance_id": "237-126133-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0007.flac", "task_type": "lexical_detection", "query": "stool", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"stool\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.51, 7.02, "stool"]], "response_mode": "offline", "answer": "[{\"word\":\"stool\",\"start\":6.51,\"end\":7.02}]"} {"utterance_id": "8463-294828-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0013.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.74, 4.11, "third"]], "response_mode": "offline", "answer": "[{\"word\":\"third\",\"start\":3.74,\"end\":4.11}]"} {"utterance_id": "8555-292519-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0001.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "Listen to the full audio. Count every occurrence of a plant name. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.82, 10.84, "one"], [11.93, 12.49, "two"], [15.48, 16.01, "three"], [19.22, 20.33, "four"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":9.82,\"end\":10.84},{\"word\":\"two\",\"start\":11.93,\"end\":12.49},{\"word\":\"three\",\"start\":15.48,\"end\":16.01},{\"word\":\"four\",\"start\":19.22,\"end\":20.33}]"} {"utterance_id": "1580-141083-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0028.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Do not answer until the audio ends. Then list each instance of an ordinal number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.88, 2.37, "next"]], "response_mode": "offline", "answer": "[{\"word\":\"next\",\"start\":1.88,\"end\":2.37}]"} {"utterance_id": "2830-3980-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0008.flac", "task_type": "lexical_detection", "query": "takes", "prompt": "Process the full audio offline. Find every instance of the word \"takes\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.59, 0.86, "takes"]], "response_mode": "offline", "answer": "[{\"word\":\"takes\",\"start\":0.59,\"end\":0.86}]"} {"utterance_id": "237-126133-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.54, 1.12, "phronsie"], [2.57, 3.05, "polly's"]], "response_mode": "offline", "answer": "[{\"word\":\"phronsie\",\"start\":0.54,\"end\":1.12},{\"word\":\"polly's\",\"start\":2.57,\"end\":3.05}]"} {"utterance_id": "4077-13754-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0010.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After the audio ends, extract all occurrences of an organization name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.63, 8.02, "senate"]], "response_mode": "offline", "answer": "[{\"word\":\"senate\",\"start\":7.63,\"end\":8.02}]"} {"utterance_id": "4446-2271-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0021.flac", "task_type": "lexical_detection", "query": "theatre", "prompt": "After listening to the whole recording, output every instance of the word \"theatre\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.28, 1.73, "theatre"]], "response_mode": "offline", "answer": "[{\"word\":\"theatre\",\"start\":1.28,\"end\":1.73}]"} {"utterance_id": "2830-3980-0065", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0065.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.32, 0.6, "paul"], [2.57, 3.5, "jesus christ"]], "response_mode": "offline", "answer": "[{\"word\":\"paul\",\"start\":0.32,\"end\":0.6},{\"word\":\"jesus christ\",\"start\":2.57,\"end\":3.5}]"} {"utterance_id": "4992-41806-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0003.flac", "task_type": "lexical_detection", "query": "recited", "prompt": "Once the audio has finished, identify every instance of the word \"recited\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.06, 3.58, "recited"]], "response_mode": "offline", "answer": "[{\"word\":\"recited\",\"start\":3.06,\"end\":3.58}]"} {"utterance_id": "7176-92135-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0043.flac", "task_type": "lexical_detection", "query": "actor", "prompt": "Once the audio has finished, identify every instance of the word \"actor\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[11.7, 11.99, "actor"]], "response_mode": "offline", "answer": "[{\"word\":\"actor\",\"start\":11.7,\"end\":11.99}]"} {"utterance_id": "121-121726-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0012.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen to the full audio. After it finishes, find every instance of a money or currency term. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.06, 2.63, "bond"]], "response_mode": "offline", "answer": "[{\"word\":\"bond\",\"start\":2.06,\"end\":2.63}]"} {"utterance_id": "7021-85628-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0000.flac", "task_type": "lexical_detection", "query": "anders", "prompt": "After listening to the whole recording, output every instance of the word \"anders\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.67, 1.17, "anders"]], "response_mode": "offline", "answer": "[{\"word\":\"anders\",\"start\":0.67,\"end\":1.17}]"} {"utterance_id": "3575-170457-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0039.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.48, 2.74, "anne"]], "response_mode": "offline", "answer": "[{\"word\":\"anne\",\"start\":2.48,\"end\":2.74}]"} {"utterance_id": "6829-68771-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0033.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[1.38, 1.77, "feet"]], "response_mode": "offline", "answer": "[{\"word\":\"feet\",\"start\":1.38,\"end\":1.77}]"} {"utterance_id": "1089-134691-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0011.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Once the audio has finished, list each occurrence of a body part as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[3.68, 4.25, "one"], [9.54, 10.04, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.68,\"end\":4.25},{\"word\":\"two\",\"start\":9.54,\"end\":10.04}]"} {"utterance_id": "5142-36377-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0008.flac", "task_type": "lexical_detection", "query": "meadowcroft", "prompt": "After listening to the whole recording, output every instance of the word \"meadowcroft\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.49, 1.01, "meadowcroft"]], "response_mode": "offline", "answer": "[{\"word\":\"meadowcroft\",\"start\":0.49,\"end\":1.01}]"} {"utterance_id": "1320-122612-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0001.flac", "task_type": "lexical_detection", "query": "dews", "prompt": "When the utterance is complete, report all instances of the word \"dews\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 0.79, "dews"]], "response_mode": "offline", "answer": "[{\"word\":\"dews\",\"start\":0.42,\"end\":0.79}]"} {"utterance_id": "1320-122617-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0018.flac", "task_type": "lexical_detection", "query": "dying", "prompt": "After listening to the whole recording, output every instance of the word \"dying\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.71, 6.11, "dying"]], "response_mode": "offline", "answer": "[{\"word\":\"dying\",\"start\":5.71,\"end\":6.11}]"} {"utterance_id": "3729-6852-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0017.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.41, 12.92, "waiters"]], "response_mode": "offline", "answer": "[{\"word\":\"waiters\",\"start\":12.41,\"end\":12.92}]"} {"utterance_id": "7729-102255-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0023.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[2.98, 3.42, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.98,\"end\":3.42}]"} {"utterance_id": "5105-28240-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0022.flac", "task_type": "semantic_detection", "query": "date", "prompt": "Listen carefully to the entire clip, then report every instance of a date as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.71, 2.31, "january"]], "response_mode": "offline", "answer": "[{\"word\":\"january\",\"start\":1.71,\"end\":2.31}]"} {"utterance_id": "1221-135767-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0003.flac", "task_type": "semantic_counting", "query": "organization", "prompt": "When the utterance is complete, report how an organization name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.73, 11.88, "one"], [17.3, 18.12, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":10.73,\"end\":11.88},{\"word\":\"two\",\"start\":17.3,\"end\":18.12}]"} {"utterance_id": "8555-292519-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0003.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.32, 8.83, "king's"]], "response_mode": "offline", "answer": "[{\"word\":\"king's\",\"start\":8.32,\"end\":8.83}]"} {"utterance_id": "2961-960-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.99, 2.58, "plato"]], "response_mode": "offline", "answer": "[{\"word\":\"plato\",\"start\":1.99,\"end\":2.58}]"} {"utterance_id": "5142-33396-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0018.flac", "task_type": "lexical_detection", "query": "belly", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"belly\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.78, 1.12, "belly"]], "response_mode": "offline", "answer": "[{\"word\":\"belly\",\"start\":0.78,\"end\":1.12}]"} {"utterance_id": "5639-40744-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0006.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.06, 5.61, "parents"]], "response_mode": "offline", "answer": "[{\"word\":\"parents\",\"start\":5.06,\"end\":5.61}]"} {"utterance_id": "4446-2275-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0004.flac", "task_type": "lexical_detection", "query": "sit", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"sit\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.45, 1.65, "sit"]], "response_mode": "offline", "answer": "[{\"word\":\"sit\",\"start\":1.45,\"end\":1.65}]"} {"utterance_id": "237-126133-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0021.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.39, 3.84, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":3.39,\"end\":3.84}]"} {"utterance_id": "1995-1837-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0003.flac", "task_type": "lexical_detection", "query": "brightened", "prompt": "Process the full audio offline. Find every instance of the word \"brightened\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.22, 2.69, "brightened"]], "response_mode": "offline", "answer": "[{\"word\":\"brightened\",\"start\":2.22,\"end\":2.69}]"} {"utterance_id": "6829-68771-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0010.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Do not answer until the audio ends. Then list each instance of an organization name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.97, 1.35, "band"]], "response_mode": "offline", "answer": "[{\"word\":\"band\",\"start\":0.97,\"end\":1.35}]"} {"utterance_id": "1580-141084-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0002.flac", "task_type": "lexical_detection", "query": "rooms", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"rooms\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.69, 1.03, "rooms"]], "response_mode": "offline", "answer": "[{\"word\":\"rooms\",\"start\":0.69,\"end\":1.03}]"} {"utterance_id": "2094-142345-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0038.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.13, 14.5, "child"]], "response_mode": "offline", "answer": "[{\"word\":\"child\",\"start\":14.13,\"end\":14.5}]"} {"utterance_id": "237-134500-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0033.flac", "task_type": "lexical_detection", "query": "weren't", "prompt": "After the audio ends, extract all occurrences of the word \"weren't\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.95, 1.21, "weren't"]], "response_mode": "offline", "answer": "[{\"word\":\"weren't\",\"start\":0.95,\"end\":1.21}]"} {"utterance_id": "4507-16021-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0026.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[15.09, 15.51, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":15.09,\"end\":15.51}]"} {"utterance_id": "1188-133604-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0026.flac", "task_type": "semantic_detection", "query": "color", "prompt": "You will hear a complete utterance. When it ends, list every instance of a color name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.0, 5.39, "golden"]], "response_mode": "offline", "answer": "[{\"word\":\"golden\",\"start\":5.0,\"end\":5.39}]"} {"utterance_id": "7729-102255-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0044.flac", "task_type": "lexical_detection", "query": "martin", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"martin\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.78, 3.39, "martin"]], "response_mode": "offline", "answer": "[{\"word\":\"martin\",\"start\":2.78,\"end\":3.39}]"} {"utterance_id": "4970-29093-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0018.flac", "task_type": "lexical_detection", "query": "wall", "prompt": "Process the full audio offline. Find every instance of the word \"wall\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.02, 5.27, "wall"]], "response_mode": "offline", "answer": "[{\"word\":\"wall\",\"start\":5.02,\"end\":5.27}]"} {"utterance_id": "1320-122612-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0004.flac", "task_type": "lexical_detection", "query": "chingachgook", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"chingachgook\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.7, 4.4, "chingachgook"]], "response_mode": "offline", "answer": "[{\"word\":\"chingachgook\",\"start\":3.7,\"end\":4.4}]"} {"utterance_id": "1320-122617-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0019.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the full audio. After it finishes, find every instance of a plant name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.38, 7.88, "withes"]], "response_mode": "offline", "answer": "[{\"word\":\"withes\",\"start\":7.38,\"end\":7.88}]"} {"utterance_id": "8455-210777-0049", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0049.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.48, 1.93, "lieutenant crosstrees"]], "response_mode": "offline", "answer": "[{\"word\":\"lieutenant crosstrees\",\"start\":0.48,\"end\":1.93}]"} {"utterance_id": "6829-68771-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0012.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a city name with fields word, start, end (seconds). If none, reply [].", "labels": [[5.64, 6.34, "elmhurst"]], "response_mode": "offline", "answer": "[{\"word\":\"elmhurst\",\"start\":5.64,\"end\":6.34}]"} {"utterance_id": "5683-32866-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0030.flac", "task_type": "lexical_detection", "query": "startled", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"startled\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.98, 3.38, "startled"]], "response_mode": "offline", "answer": "[{\"word\":\"startled\",\"start\":2.98,\"end\":3.38}]"} {"utterance_id": "4970-29095-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0022.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.44, 3.61, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.44,\"end\":3.61}]"} {"utterance_id": "1284-1181-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0018.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.38, 2.99, "magician"]], "response_mode": "offline", "answer": "[{\"word\":\"magician\",\"start\":2.38,\"end\":2.99}]"} {"utterance_id": "5142-36377-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0024.flac", "task_type": "lexical_counting", "query": "two", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"two\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[3.24, 3.55, "one"], [4.9, 5.1, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.24,\"end\":3.55},{\"word\":\"two\",\"start\":4.9,\"end\":5.1}]"} {"utterance_id": "1221-135767-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0013.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[5.69, 5.9, "one"], [10.2, 10.6, "seven"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.69,\"end\":5.9},{\"word\":\"seven\",\"start\":10.2,\"end\":10.6}]"} {"utterance_id": "1580-141083-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0012.flac", "task_type": "semantic_detection", "query": "color", "prompt": "When the utterance is complete, report all instances of a color name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.98, 3.44, "black"]], "response_mode": "offline", "answer": "[{\"word\":\"black\",\"start\":2.98,\"end\":3.44}]"} {"utterance_id": "1188-133604-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0026.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "After the audio ends, tally every instance of a person's name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[2.91, 3.5, "one"], [18.2, 18.74, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.91,\"end\":3.5},{\"word\":\"two\",\"start\":18.2,\"end\":18.74}]"} {"utterance_id": "8555-284447-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0000.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.34, 4.82, "soldiers"], [7.48, 7.95, "sailorman"]], "response_mode": "offline", "answer": "[{\"word\":\"soldiers\",\"start\":4.34,\"end\":4.82},{\"word\":\"sailorman\",\"start\":7.48,\"end\":7.95}]"} {"utterance_id": "8455-210777-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0038.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a profession (word, start, end). If none, reply [].", "labels": [[3.77, 4.36, "one"], [7.77, 8.37, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.77,\"end\":4.36},{\"word\":\"two\",\"start\":7.77,\"end\":8.37}]"} {"utterance_id": "4507-16021-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0059.flac", "task_type": "lexical_detection", "query": "marvel", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"marvel\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.55, 5.04, "marvel"]], "response_mode": "offline", "answer": "[{\"word\":\"marvel\",\"start\":4.55,\"end\":5.04}]"} {"utterance_id": "4446-2275-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0026.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.06, 1.41, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":1.06,\"end\":1.41}]"} {"utterance_id": "6829-68769-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0048.flac", "task_type": "semantic_counting", "query": "money", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a money or currency term: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.68, 0.99, "one"], [4.19, 4.51, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.68,\"end\":0.99},{\"word\":\"two\",\"start\":4.19,\"end\":4.51}]"} {"utterance_id": "2961-961-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0007.flac", "task_type": "lexical_detection", "query": "described", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"described\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.26, 3.98, "described"]], "response_mode": "offline", "answer": "[{\"word\":\"described\",\"start\":3.26,\"end\":3.98}]"} {"utterance_id": "908-157963-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0013.flac", "task_type": "lexical_detection", "query": "air", "prompt": "Process the full audio offline. Find every instance of the word \"air\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.57, 4.02, "air"]], "response_mode": "offline", "answer": "[{\"word\":\"air\",\"start\":3.57,\"end\":4.02}]"} {"utterance_id": "3575-170457-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0025.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.79, 7.11, "thirty"]], "response_mode": "offline", "answer": "[{\"word\":\"thirty\",\"start\":6.79,\"end\":7.11}]"} {"utterance_id": "7729-102255-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0004.flac", "task_type": "semantic_counting", "query": "state_or_province", "prompt": "Listen to the full audio. Count every occurrence of a state or province name. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.63, 1.04, "one"], [7.06, 7.49, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.63,\"end\":1.04},{\"word\":\"two\",\"start\":7.06,\"end\":7.49}]"} {"utterance_id": "7176-92135-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0015.flac", "task_type": "lexical_detection", "query": "non", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"non\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.22, 5.5, "non"]], "response_mode": "offline", "answer": "[{\"word\":\"non\",\"start\":5.22,\"end\":5.5}]"} {"utterance_id": "237-134500-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0028.flac", "task_type": "lexical_detection", "query": "murmured", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"murmured\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.65, 4.12, "murmured"]], "response_mode": "offline", "answer": "[{\"word\":\"murmured\",\"start\":3.65,\"end\":4.12}]"} {"utterance_id": "7176-92135-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0012.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.77, 10.11, "jaw"]], "response_mode": "offline", "answer": "[{\"word\":\"jaw\",\"start\":9.77,\"end\":10.11}]"} {"utterance_id": "1221-135766-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0006.flac", "task_type": "lexical_detection", "query": "disposition", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"disposition\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.79, 4.61, "disposition"]], "response_mode": "offline", "answer": "[{\"word\":\"disposition\",\"start\":3.79,\"end\":4.61}]"} {"utterance_id": "7127-75947-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0004.flac", "task_type": "lexical_detection", "query": "explain", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"explain\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.53, 1.0, "explain"]], "response_mode": "offline", "answer": "[{\"word\":\"explain\",\"start\":0.53,\"end\":1.0}]"} {"utterance_id": "1221-135767-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0021.flac", "task_type": "lexical_detection", "query": "bow", "prompt": "When the utterance is complete, report all instances of the word \"bow\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.78, 1.99, "bow"]], "response_mode": "offline", "answer": "[{\"word\":\"bow\",\"start\":1.78,\"end\":1.99}]"} {"utterance_id": "1580-141083-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0008.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.88, 1.27, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":0.88,\"end\":1.27}]"} {"utterance_id": "1995-1826-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0013.flac", "task_type": "lexical_detection", "query": "barn", "prompt": "When the utterance is complete, report all instances of the word \"barn\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.72, 6.25, "barn"]], "response_mode": "offline", "answer": "[{\"word\":\"barn\",\"start\":5.72,\"end\":6.25}]"} {"utterance_id": "4077-13751-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0006.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.61, 0.86, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":0.61,\"end\":0.86}]"} {"utterance_id": "1089-134686-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0008.flac", "task_type": "lexical_detection", "query": "extinguished", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"extinguished\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.27, 2.96, "extinguished"]], "response_mode": "offline", "answer": "[{\"word\":\"extinguished\",\"start\":2.27,\"end\":2.96}]"} {"utterance_id": "2830-3980-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.35, 0.59, "paul"]], "response_mode": "offline", "answer": "[{\"word\":\"paul\",\"start\":0.35,\"end\":0.59}]"} {"utterance_id": "237-134500-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0000.flac", "task_type": "lexical_detection", "query": "frank", "prompt": "After the audio ends, extract all occurrences of the word \"frank\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.33, 0.71, "frank"]], "response_mode": "offline", "answer": "[{\"word\":\"frank\",\"start\":0.33,\"end\":0.71}]"} {"utterance_id": "2300-131720-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0012.flac", "task_type": "lexical_detection", "query": "test", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"test\" with fields word, start, end (seconds). If none, reply [].", "labels": [[8.93, 9.28, "test"]], "response_mode": "offline", "answer": "[{\"word\":\"test\",\"start\":8.93,\"end\":9.28}]"} {"utterance_id": "7729-102255-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0007.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen carefully to the entire clip, then report every instance of a city name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.59, 2.26, "westport"]], "response_mode": "offline", "answer": "[{\"word\":\"westport\",\"start\":1.59,\"end\":2.26}]"} {"utterance_id": "5683-32865-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0003.flac", "task_type": "lexical_counting", "query": "cousins", "prompt": "Listen to the full audio. Count every occurrence of the word \"cousins\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.66, 1.22, "one"], [2.61, 3.33, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.66,\"end\":1.22},{\"word\":\"two\",\"start\":2.61,\"end\":3.33}]"} {"utterance_id": "908-31957-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0015.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Process the full audio offline. Find every instance of a color name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.51, 12.13, "purple"]], "response_mode": "offline", "answer": "[{\"word\":\"purple\",\"start\":11.51,\"end\":12.13}]"} {"utterance_id": "7127-75946-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0020.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.9, 5.5, "musicians"]], "response_mode": "offline", "answer": "[{\"word\":\"musicians\",\"start\":4.9,\"end\":5.5}]"} {"utterance_id": "4446-2273-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0032.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "When the utterance is complete, report all instances of a weather term including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.0, 5.48, "misty"]], "response_mode": "offline", "answer": "[{\"word\":\"misty\",\"start\":5.0,\"end\":5.48}]"} {"utterance_id": "8555-292519-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0000.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "When the utterance is complete, report all instances of a time of day including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.66, 2.18, "dawn's"]], "response_mode": "offline", "answer": "[{\"word\":\"dawn's\",\"start\":1.66,\"end\":2.18}]"} {"utterance_id": "4992-41797-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0021.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.55, 7.02, "teacher"]], "response_mode": "offline", "answer": "[{\"word\":\"teacher\",\"start\":6.55,\"end\":7.02}]"} {"utterance_id": "8230-279154-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0023.flac", "task_type": "lexical_detection", "query": "vague", "prompt": "When the utterance is complete, report all instances of the word \"vague\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.58, 6.02, "vague"]], "response_mode": "offline", "answer": "[{\"word\":\"vague\",\"start\":5.58,\"end\":6.02}]"} {"utterance_id": "4446-2273-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0003.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen carefully to the entire clip, then report every instance of a time of day as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.77, 3.16, "evening"]], "response_mode": "offline", "answer": "[{\"word\":\"evening\",\"start\":2.77,\"end\":3.16}]"} {"utterance_id": "7127-75947-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0030.flac", "task_type": "lexical_detection", "query": "count", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"count\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.05, 2.45, "count"]], "response_mode": "offline", "answer": "[{\"word\":\"count\",\"start\":2.05,\"end\":2.45}]"} {"utterance_id": "4970-29095-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0035.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Once the audio has finished, identify every instance of a money or currency term. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.14, 1.74, "fortune"]], "response_mode": "offline", "answer": "[{\"word\":\"fortune\",\"start\":1.14,\"end\":1.74}]"} {"utterance_id": "6829-68771-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0001.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen carefully to the entire clip, then report every instance of an ordinal number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.57, 1.88, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":1.57,\"end\":1.88}]"} {"utterance_id": "1320-122617-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0015.flac", "task_type": "lexical_detection", "query": "obeying", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"obeying\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.38, 1.95, "obeying"]], "response_mode": "offline", "answer": "[{\"word\":\"obeying\",\"start\":1.38,\"end\":1.95}]"} {"utterance_id": "1995-1826-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0020.flac", "task_type": "lexical_detection", "query": "remark", "prompt": "After listening to the whole recording, output every instance of the word \"remark\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.28, 4.79, "remark"]], "response_mode": "offline", "answer": "[{\"word\":\"remark\",\"start\":4.28,\"end\":4.79}]"} {"utterance_id": "2830-3979-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0011.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.63, 9.04, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":8.63,\"end\":9.04}]"} {"utterance_id": "2961-960-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0001.flac", "task_type": "lexical_detection", "query": "posterity", "prompt": "After listening to the whole recording, output every instance of the word \"posterity\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.14, 4.97, "posterity"]], "response_mode": "offline", "answer": "[{\"word\":\"posterity\",\"start\":4.14,\"end\":4.97}]"} {"utterance_id": "672-122797-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0003.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Once the audio has finished, identify every instance of a plant name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.41, 2.73, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":2.41,\"end\":2.73}]"} {"utterance_id": "2094-142345-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0010.flac", "task_type": "lexical_counting", "query": "turned", "prompt": "Listen to the full audio. Count every occurrence of the word \"turned\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.62, 5.07, "one"], [12.73, 13.13, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.62,\"end\":5.07},{\"word\":\"two\",\"start\":12.73,\"end\":13.13}]"} {"utterance_id": "4507-16021-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0015.flac", "task_type": "lexical_detection", "query": "medicine", "prompt": "After listening to the whole recording, output every instance of the word \"medicine\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.73, 3.49, "medicine"]], "response_mode": "offline", "answer": "[{\"word\":\"medicine\",\"start\":2.73,\"end\":3.49}]"} {"utterance_id": "8224-274384-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0013.flac", "task_type": "lexical_detection", "query": "misfortune", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"misfortune\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.25, 5.14, "misfortune"]], "response_mode": "offline", "answer": "[{\"word\":\"misfortune\",\"start\":4.25,\"end\":5.14}]"} {"utterance_id": "8224-274381-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0012.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the full audio. Count every occurrence of a person's name. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.33, 1.22, "one"], [16.92, 17.81, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.33,\"end\":1.22},{\"word\":\"two\",\"start\":16.92,\"end\":17.81}]"} {"utterance_id": "6930-75918-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0010.flac", "task_type": "lexical_detection", "query": "clearly", "prompt": "Process the full audio offline. Find every instance of the word \"clearly\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.88, 2.3, "clearly"]], "response_mode": "offline", "answer": "[{\"word\":\"clearly\",\"start\":1.88,\"end\":2.3}]"} {"utterance_id": "6829-68769-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0011.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.79, 1.55, "representative"]], "response_mode": "offline", "answer": "[{\"word\":\"representative\",\"start\":0.79,\"end\":1.55}]"} {"utterance_id": "3570-5694-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0005.flac", "task_type": "lexical_detection", "query": "particularly", "prompt": "After listening to the whole recording, output every instance of the word \"particularly\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.71, 3.45, "particularly"]], "response_mode": "offline", "answer": "[{\"word\":\"particularly\",\"start\":2.71,\"end\":3.45}]"} {"utterance_id": "5142-33396-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0052.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.99, 1.24, "sif"]], "response_mode": "offline", "answer": "[{\"word\":\"sif\",\"start\":0.99,\"end\":1.24}]"} {"utterance_id": "5142-36377-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0021.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "When the utterance is complete, report how a person's name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.79, 11.38, "one"], [15.92, 16.5, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":10.79,\"end\":11.38},{\"word\":\"two\",\"start\":15.92,\"end\":16.5}]"} {"utterance_id": "5639-40744-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0014.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.67, 3.02, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":2.67,\"end\":3.02}]"} {"utterance_id": "7729-102255-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0042.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Once the audio has finished, identify every instance of a state or province name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[8.56, 9.15, "missouri"]], "response_mode": "offline", "answer": "[{\"word\":\"missouri\",\"start\":8.56,\"end\":9.15}]"} {"utterance_id": "2830-3980-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0005.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the full audio. After it finishes, find every instance of an organization name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.49, 6.06, "church"]], "response_mode": "offline", "answer": "[{\"word\":\"church\",\"start\":5.49,\"end\":6.06}]"} {"utterance_id": "2094-142345-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0036.flac", "task_type": "lexical_detection", "query": "families", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"families\" with fields word, start, end (seconds). If none, reply [].", "labels": [[5.75, 6.47, "families"]], "response_mode": "offline", "answer": "[{\"word\":\"families\",\"start\":5.75,\"end\":6.47}]"} {"utterance_id": "4446-2273-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.3, 0.7, "hilda"]], "response_mode": "offline", "answer": "[{\"word\":\"hilda\",\"start\":0.3,\"end\":0.7}]"} {"utterance_id": "7127-75946-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0026.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "After listening to the whole recording, output timed count words for a profession (one, two, three, ...) as JSON. If none, reply [].", "labels": [[4.41, 5.09, "one"], [12.95, 13.42, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.41,\"end\":5.09},{\"word\":\"two\",\"start\":12.95,\"end\":13.42}]"} {"utterance_id": "4992-41797-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0004.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.65, 1.33, "swan"]], "response_mode": "offline", "answer": "[{\"word\":\"swan\",\"start\":0.65,\"end\":1.33}]"} {"utterance_id": "121-127105-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0033.flac", "task_type": "lexical_detection", "query": "beauty", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"beauty\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.32, 1.75, "beauty"]], "response_mode": "offline", "answer": "[{\"word\":\"beauty\",\"start\":1.32,\"end\":1.75}]"} {"utterance_id": "1320-122617-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0023.flac", "task_type": "lexical_detection", "query": "recoiled", "prompt": "When the utterance is complete, report all instances of the word \"recoiled\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.24, 4.91, "recoiled"]], "response_mode": "offline", "answer": "[{\"word\":\"recoiled\",\"start\":4.24,\"end\":4.91}]"} {"utterance_id": "1089-134686-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0012.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "After listening to the whole recording, output timed count words for a body part (one, two, three, ...) as JSON. If none, reply [].", "labels": [[0.7, 1.06, "one"], [7.77, 8.31, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.7,\"end\":1.06},{\"word\":\"two\",\"start\":7.77,\"end\":8.31}]"} {"utterance_id": "1089-134686-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0011.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an organization name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.99, 2.63, "sodality"]], "response_mode": "offline", "answer": "[{\"word\":\"sodality\",\"start\":1.99,\"end\":2.63}]"} {"utterance_id": "121-127105-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0013.flac", "task_type": "lexical_detection", "query": "fix", "prompt": "After listening to the whole recording, output every instance of the word \"fix\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.79, 5.1, "fix"]], "response_mode": "offline", "answer": "[{\"word\":\"fix\",\"start\":4.79,\"end\":5.1}]"} {"utterance_id": "3570-5695-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0005.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.88, 11.39, "children"]], "response_mode": "offline", "answer": "[{\"word\":\"children\",\"start\":10.88,\"end\":11.39}]"} {"utterance_id": "1089-134686-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0025.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.96, 4.44, "stephen"]], "response_mode": "offline", "answer": "[{\"word\":\"stephen\",\"start\":3.96,\"end\":4.44}]"} {"utterance_id": "1320-122612-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0001.flac", "task_type": "semantic_counting", "query": "weather", "prompt": "Listen to the full audio. Count every occurrence of a weather term. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 0.79, "one"], [3.62, 4.1, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.42,\"end\":0.79},{\"word\":\"two\",\"start\":3.62,\"end\":4.1}]"} {"utterance_id": "5142-33396-0064", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0064.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the full audio. After it finishes, find every instance of a family relation. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.48, 0.85, "father"], [2.06, 2.32, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":0.48,\"end\":0.85},{\"word\":\"mother\",\"start\":2.06,\"end\":2.32}]"} {"utterance_id": "8555-284449-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0005.flac", "task_type": "lexical_detection", "query": "finished", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"finished\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.61, 1.15, "finished"]], "response_mode": "offline", "answer": "[{\"word\":\"finished\",\"start\":0.61,\"end\":1.15}]"} {"utterance_id": "4446-2273-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0033.flac", "task_type": "lexical_detection", "query": "neither", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"neither\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.36, 1.54, "neither"]], "response_mode": "offline", "answer": "[{\"word\":\"neither\",\"start\":1.36,\"end\":1.54}]"} {"utterance_id": "672-122797-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0038.flac", "task_type": "lexical_detection", "query": "believed", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"believed\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.17, 2.55, "believed"]], "response_mode": "offline", "answer": "[{\"word\":\"believed\",\"start\":2.17,\"end\":2.55}]"} {"utterance_id": "5683-32879-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0004.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Once the audio has finished, identify every instance of a color name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[8.29, 8.63, "pale"]], "response_mode": "offline", "answer": "[{\"word\":\"pale\",\"start\":8.29,\"end\":8.63}]"} {"utterance_id": "8455-210777-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0044.flac", "task_type": "lexical_detection", "query": "carried", "prompt": "Once the audio has finished, identify every instance of the word \"carried\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.83, 2.18, "carried"]], "response_mode": "offline", "answer": "[{\"word\":\"carried\",\"start\":1.83,\"end\":2.18}]"} {"utterance_id": "1221-135766-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0012.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.57, 2.87, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":2.57,\"end\":2.87}]"} {"utterance_id": "8463-287645-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0011.flac", "task_type": "lexical_detection", "query": "community", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"community\" with fields word, start, end (seconds). If none, reply [].", "labels": [[5.54, 6.23, "community"]], "response_mode": "offline", "answer": "[{\"word\":\"community\",\"start\":5.54,\"end\":6.23}]"} {"utterance_id": "1188-133604-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0008.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen carefully to the entire clip, then report every instance of an animal name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.81, 6.51, "cock robin"]], "response_mode": "offline", "answer": "[{\"word\":\"cock robin\",\"start\":5.81,\"end\":6.51}]"} {"utterance_id": "4077-13754-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0012.flac", "task_type": "lexical_detection", "query": "alleged", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"alleged\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.11, 2.57, "alleged"]], "response_mode": "offline", "answer": "[{\"word\":\"alleged\",\"start\":2.11,\"end\":2.57}]"} {"utterance_id": "1995-1836-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0001.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "You will hear a complete utterance. When it ends, list every instance of a plant name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.9, 1.2, "cotton"]], "response_mode": "offline", "answer": "[{\"word\":\"cotton\",\"start\":0.9,\"end\":1.2}]"} {"utterance_id": "2830-3980-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0052.flac", "task_type": "lexical_detection", "query": "divine", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"divine\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.63, 4.01, "divine"]], "response_mode": "offline", "answer": "[{\"word\":\"divine\",\"start\":3.63,\"end\":4.01}]"} {"utterance_id": "1320-122612-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0013.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.28, 1.62, "hundred"]], "response_mode": "offline", "answer": "[{\"word\":\"hundred\",\"start\":1.28,\"end\":1.62}]"} {"utterance_id": "7127-75946-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0013.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Process the full audio offline. Find every instance of a time of day and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.46, 3.24, "nine o'clock"]], "response_mode": "offline", "answer": "[{\"word\":\"nine o'clock\",\"start\":2.46,\"end\":3.24}]"} {"utterance_id": "260-123286-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0006.flac", "task_type": "lexical_detection", "query": "losing", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"losing\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.49, 0.79, "losing"]], "response_mode": "offline", "answer": "[{\"word\":\"losing\",\"start\":0.49,\"end\":0.79}]"} {"utterance_id": "4992-23283-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0003.flac", "task_type": "lexical_detection", "query": "added", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"added\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.2, 2.46, "added"]], "response_mode": "offline", "answer": "[{\"word\":\"added\",\"start\":2.2,\"end\":2.46}]"} {"utterance_id": "1320-122617-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0002.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.59, 13.23, "organs"]], "response_mode": "offline", "answer": "[{\"word\":\"organs\",\"start\":12.59,\"end\":13.23}]"} {"utterance_id": "7176-88083-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0026.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.21, 1.58, "beak"], [2.85, 3.18, "wings"]], "response_mode": "offline", "answer": "[{\"word\":\"beak\",\"start\":1.21,\"end\":1.58},{\"word\":\"wings\",\"start\":2.85,\"end\":3.18}]"} {"utterance_id": "7176-92135-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0016.flac", "task_type": "lexical_counting", "query": "easy", "prompt": "Process the full audio offline. For every occurrence of the word \"easy\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.7, 1.02, "one"], [5.56, 5.86, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.7,\"end\":1.02},{\"word\":\"two\",\"start\":5.56,\"end\":5.86}]"} {"utterance_id": "1284-1181-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0003.flac", "task_type": "lexical_detection", "query": "yarn", "prompt": "When the utterance is complete, report all instances of the word \"yarn\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.38, 1.81, "yarn"]], "response_mode": "offline", "answer": "[{\"word\":\"yarn\",\"start\":1.38,\"end\":1.81}]"} {"utterance_id": "2094-142345-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0015.flac", "task_type": "semantic_detection", "query": "food", "prompt": "You will hear a complete utterance. When it ends, list every instance of a food name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.44, 3.71, "dinner"]], "response_mode": "offline", "answer": "[{\"word\":\"dinner\",\"start\":3.44,\"end\":3.71}]"} {"utterance_id": "61-70968-0061", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0061.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[1.4, 1.81, "leech"]], "response_mode": "offline", "answer": "[{\"word\":\"leech\",\"start\":1.4,\"end\":1.81}]"} {"utterance_id": "61-70970-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0007.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.87, 2.25, "clerk"]], "response_mode": "offline", "answer": "[{\"word\":\"clerk\",\"start\":1.87,\"end\":2.25}]"} {"utterance_id": "237-126133-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0024.flac", "task_type": "lexical_detection", "query": "laughter", "prompt": "When the utterance is complete, report all instances of the word \"laughter\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.58, 7.09, "laughter"]], "response_mode": "offline", "answer": "[{\"word\":\"laughter\",\"start\":6.58,\"end\":7.09}]"} {"utterance_id": "1320-122617-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0013.flac", "task_type": "lexical_detection", "query": "delivered", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"delivered\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.15, 0.62, "delivered"]], "response_mode": "offline", "answer": "[{\"word\":\"delivered\",\"start\":0.15,\"end\":0.62}]"} {"utterance_id": "4446-2271-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0023.flac", "task_type": "lexical_detection", "query": "awfully", "prompt": "After listening to the whole recording, output every instance of the word \"awfully\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.22, 1.66, "awfully"]], "response_mode": "offline", "answer": "[{\"word\":\"awfully\",\"start\":1.22,\"end\":1.66}]"} {"utterance_id": "8224-274381-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0015.flac", "task_type": "lexical_detection", "query": "discipline", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"discipline\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.86, 1.44, "discipline"]], "response_mode": "offline", "answer": "[{\"word\":\"discipline\",\"start\":0.86,\"end\":1.44}]"} {"utterance_id": "5142-33396-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0033.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.46, 2.73, "half"]], "response_mode": "offline", "answer": "[{\"word\":\"half\",\"start\":2.46,\"end\":2.73}]"} {"utterance_id": "1580-141084-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0014.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.56, 1.22, "bannister"]], "response_mode": "offline", "answer": "[{\"word\":\"bannister\",\"start\":0.56,\"end\":1.22}]"} {"utterance_id": "5142-36377-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0021.flac", "task_type": "lexical_counting", "query": "interest", "prompt": "Once the audio has finished, list each occurrence of the word \"interest\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[10.32, 10.73, "one"], [11.94, 12.29, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":10.32,\"end\":10.73},{\"word\":\"two\",\"start\":11.94,\"end\":12.29}]"} {"utterance_id": "5683-32866-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0008.flac", "task_type": "lexical_detection", "query": "fellow", "prompt": "Process the full audio offline. Find every instance of the word \"fellow\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.32, 1.78, "fellow"]], "response_mode": "offline", "answer": "[{\"word\":\"fellow\",\"start\":1.32,\"end\":1.78}]"} {"utterance_id": "237-126133-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0001.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.69, 9.99, "wings"]], "response_mode": "offline", "answer": "[{\"word\":\"wings\",\"start\":9.69,\"end\":9.99}]"} {"utterance_id": "1284-1180-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0028.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the full audio. After it finishes, find every instance of a color name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.29, 3.7, "colors"]], "response_mode": "offline", "answer": "[{\"word\":\"colors\",\"start\":3.29,\"end\":3.7}]"} {"utterance_id": "3729-6852-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0033.flac", "task_type": "semantic_counting", "query": "ordinal", "prompt": "After listening to the whole recording, output timed count words for an ordinal number (one, two, three, ...) as JSON. If none, reply [].", "labels": [[9.81, 10.27, "one"], [15.45, 16.06, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":9.81,\"end\":10.27},{\"word\":\"two\",\"start\":15.45,\"end\":16.06}]"} {"utterance_id": "4077-13754-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0006.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.83, 10.23, "wives"]], "response_mode": "offline", "answer": "[{\"word\":\"wives\",\"start\":9.83,\"end\":10.23}]"} {"utterance_id": "4970-29095-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0020.flac", "task_type": "lexical_counting", "query": "end", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"end\" (word, start, end). If none, reply [].", "labels": [[1.7, 1.88, "one"], [7.49, 7.9, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.7,\"end\":1.88},{\"word\":\"two\",\"start\":7.49,\"end\":7.9}]"} {"utterance_id": "1221-135767-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0023.flac", "task_type": "lexical_detection", "query": "peninsula", "prompt": "When the utterance is complete, report all instances of the word \"peninsula\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.35, 10.06, "peninsula"]], "response_mode": "offline", "answer": "[{\"word\":\"peninsula\",\"start\":9.35,\"end\":10.06}]"} {"utterance_id": "8463-294825-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0002.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a country name with fields word, start, end (seconds). If none, reply [].", "labels": [[7.83, 9.24, "britain america"]], "response_mode": "offline", "answer": "[{\"word\":\"britain america\",\"start\":7.83,\"end\":9.24}]"} {"utterance_id": "260-123440-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0010.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.79, 4.34, "claws"], [7.47, 8.05, "jaws"]], "response_mode": "offline", "answer": "[{\"word\":\"claws\",\"start\":3.79,\"end\":4.34},{\"word\":\"jaws\",\"start\":7.47,\"end\":8.05}]"} {"utterance_id": "7127-75947-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0022.flac", "task_type": "lexical_detection", "query": "flatters", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"flatters\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.22, 4.72, "flatters"], [5.84, 6.3, "flatters"]], "response_mode": "offline", "answer": "[{\"word\":\"flatters\",\"start\":4.22,\"end\":4.72},{\"word\":\"flatters\",\"start\":5.84,\"end\":6.3}]"} {"utterance_id": "5105-28240-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0012.flac", "task_type": "lexical_detection", "query": "captain", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"captain\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.3, 2.74, "captain"]], "response_mode": "offline", "answer": "[{\"word\":\"captain\",\"start\":2.3,\"end\":2.74}]"} {"utterance_id": "5683-32866-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.62, 3.3, "beelzebub"]], "response_mode": "offline", "answer": "[{\"word\":\"beelzebub\",\"start\":2.62,\"end\":3.3}]"} {"utterance_id": "7021-85628-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0009.flac", "task_type": "lexical_detection", "query": "rows", "prompt": "When the utterance is complete, report all instances of the word \"rows\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.69, 8.44, "rows"]], "response_mode": "offline", "answer": "[{\"word\":\"rows\",\"start\":7.69,\"end\":8.44}]"} {"utterance_id": "237-134493-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0001.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.22, 6.5, "wheat"]], "response_mode": "offline", "answer": "[{\"word\":\"wheat\",\"start\":6.22,\"end\":6.5}]"} {"utterance_id": "2830-3980-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0024.flac", "task_type": "lexical_detection", "query": "epistle", "prompt": "When the utterance is complete, report all instances of the word \"epistle\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.15, 1.58, "epistle"]], "response_mode": "offline", "answer": "[{\"word\":\"epistle\",\"start\":1.15,\"end\":1.58}]"} {"utterance_id": "1580-141083-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0036.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.68, 2.07, "palm"]], "response_mode": "offline", "answer": "[{\"word\":\"palm\",\"start\":1.68,\"end\":2.07}]"} {"utterance_id": "5639-40744-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0006.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "After the audio ends, tally every instance of a person's name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.48, 1.13, "one"], [4.43, 4.95, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.48,\"end\":1.13},{\"word\":\"two\",\"start\":4.43,\"end\":4.95}]"} {"utterance_id": "8455-210777-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0051.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.5, 13.6, "sir ferdinando"]], "response_mode": "offline", "answer": "[{\"word\":\"sir ferdinando\",\"start\":12.5,\"end\":13.6}]"} {"utterance_id": "1580-141084-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0026.flac", "task_type": "lexical_counting", "query": "ourselves", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"ourselves\" with word/start/end. If none, reply [].", "labels": [[2.54, 3.0, "one"], [4.63, 5.15, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.54,\"end\":3.0},{\"word\":\"two\",\"start\":4.63,\"end\":5.15}]"} {"utterance_id": "7176-92135-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0041.flac", "task_type": "lexical_detection", "query": "hostess", "prompt": "After listening to the whole recording, output every instance of the word \"hostess\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.87, 4.3, "hostess"]], "response_mode": "offline", "answer": "[{\"word\":\"hostess\",\"start\":3.87,\"end\":4.3}]"} {"utterance_id": "5142-36600-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36600/5142-36600-0000.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.6, 1.15, "seven"]], "response_mode": "offline", "answer": "[{\"word\":\"seven\",\"start\":0.6,\"end\":1.15}]"} {"utterance_id": "4446-2273-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0002.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.23, 0.52, "lamb"]], "response_mode": "offline", "answer": "[{\"word\":\"lamb\",\"start\":0.23,\"end\":0.52}]"} {"utterance_id": "3575-170457-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0000.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a family relation with fields word, start, end (seconds). If none, reply [].", "labels": [[1.3, 1.63, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":1.3,\"end\":1.63}]"} {"utterance_id": "1221-135767-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0014.flac", "task_type": "lexical_detection", "query": "leech", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"leech\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.15, 6.74, "leech"]], "response_mode": "offline", "answer": "[{\"word\":\"leech\",\"start\":6.15,\"end\":6.74}]"} {"utterance_id": "672-122797-0070", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0070.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Listen carefully to the entire clip, then report every instance of a weather term as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.97, 5.74, "sunshine"]], "response_mode": "offline", "answer": "[{\"word\":\"sunshine\",\"start\":4.97,\"end\":5.74}]"} {"utterance_id": "5639-40744-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0031.flac", "task_type": "lexical_counting", "query": "rodolfo", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"rodolfo\" with word/start/end. If none, reply [].", "labels": [[16.63, 17.19, "one"], [20.66, 21.12, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":16.63,\"end\":17.19},{\"word\":\"two\",\"start\":20.66,\"end\":21.12}]"} {"utterance_id": "5639-40744-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0032.flac", "task_type": "lexical_detection", "query": "evenly", "prompt": "Once the audio has finished, identify every instance of the word \"evenly\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[8.68, 9.31, "evenly"]], "response_mode": "offline", "answer": "[{\"word\":\"evenly\",\"start\":8.68,\"end\":9.31}]"} {"utterance_id": "1188-133604-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0042.flac", "task_type": "lexical_detection", "query": "scene", "prompt": "After listening to the whole recording, output every instance of the word \"scene\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.61, 0.86, "scene"]], "response_mode": "offline", "answer": "[{\"word\":\"scene\",\"start\":0.61,\"end\":0.86}]"} {"utterance_id": "4446-2271-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0017.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[0.35, 0.6, "one"], [7.92, 8.2, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.35,\"end\":0.6},{\"word\":\"one\",\"start\":7.92,\"end\":8.2}]"} {"utterance_id": "5683-32866-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0005.flac", "task_type": "lexical_detection", "query": "hall", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"hall\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.84, 3.38, "hall"]], "response_mode": "offline", "answer": "[{\"word\":\"hall\",\"start\":2.84,\"end\":3.38}]"} {"utterance_id": "1320-122617-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0006.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.04, 2.4, "david"]], "response_mode": "offline", "answer": "[{\"word\":\"david\",\"start\":2.04,\"end\":2.4}]"} {"utterance_id": "4446-2273-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0030.flac", "task_type": "lexical_detection", "query": "alexander", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"alexander\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.39, 0.92, "alexander"]], "response_mode": "offline", "answer": "[{\"word\":\"alexander\",\"start\":0.39,\"end\":0.92}]"} {"utterance_id": "2961-960-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0011.flac", "task_type": "lexical_detection", "query": "flux", "prompt": "Once the audio has finished, identify every instance of the word \"flux\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.4, 4.14, "flux"]], "response_mode": "offline", "answer": "[{\"word\":\"flux\",\"start\":3.4,\"end\":4.14}]"} {"utterance_id": "6829-68771-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0004.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "After listening to the whole recording, output timed count words for a person's name (one, two, three, ...) as JSON. If none, reply [].", "labels": [[2.23, 2.74, "one"], [7.37, 8.01, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.23,\"end\":2.74},{\"word\":\"two\",\"start\":7.37,\"end\":8.01}]"} {"utterance_id": "5142-36377-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0022.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.29, 2.56, "hands"]], "response_mode": "offline", "answer": "[{\"word\":\"hands\",\"start\":2.29,\"end\":2.56}]"} {"utterance_id": "5639-40744-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0037.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.02, 1.35, "son"], [4.16, 4.63, "bride"], [7.36, 7.73, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"son\",\"start\":1.02,\"end\":1.35},{\"word\":\"bride\",\"start\":4.16,\"end\":4.63},{\"word\":\"father\",\"start\":7.36,\"end\":7.73}]"} {"utterance_id": "2300-131720-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0019.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.8, 10.08, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":9.8,\"end\":10.08}]"} {"utterance_id": "7127-75947-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0006.flac", "task_type": "lexical_detection", "query": "inquired", "prompt": "After the audio ends, extract all occurrences of the word \"inquired\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.74, 1.37, "inquired"]], "response_mode": "offline", "answer": "[{\"word\":\"inquired\",\"start\":0.74,\"end\":1.37}]"} {"utterance_id": "672-122797-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0022.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a plant name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.94, 3.43, "pith"], [4.46, 4.88, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"pith\",\"start\":2.94,\"end\":3.43},{\"word\":\"tree\",\"start\":4.46,\"end\":4.88}]"} {"utterance_id": "7021-85628-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0025.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.78, 1.28, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":0.78,\"end\":1.28}]"} {"utterance_id": "6930-75918-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0003.flac", "task_type": "lexical_counting", "query": "france", "prompt": "When the utterance is complete, report how the word \"france\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[15.43, 15.92, "one"], [22.33, 22.97, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":15.43,\"end\":15.92},{\"word\":\"two\",\"start\":22.33,\"end\":22.97}]"} {"utterance_id": "1284-1180-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0014.flac", "task_type": "semantic_detection", "query": "food", "prompt": "You will hear a complete utterance. When it ends, list every instance of a food name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.38, 2.68, "meal"]], "response_mode": "offline", "answer": "[{\"word\":\"meal\",\"start\":2.38,\"end\":2.68}]"} {"utterance_id": "4507-16021-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0033.flac", "task_type": "lexical_detection", "query": "mountain", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"mountain\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.71, 2.17, "mountain"]], "response_mode": "offline", "answer": "[{\"word\":\"mountain\",\"start\":1.71,\"end\":2.17}]"} {"utterance_id": "4992-23283-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0013.flac", "task_type": "lexical_detection", "query": "taste", "prompt": "After the audio ends, extract all occurrences of the word \"taste\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.34, 2.81, "taste"]], "response_mode": "offline", "answer": "[{\"word\":\"taste\",\"start\":2.34,\"end\":2.81}]"} {"utterance_id": "1995-1826-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0026.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.71, 15.18, "cotton"]], "response_mode": "offline", "answer": "[{\"word\":\"cotton\",\"start\":14.71,\"end\":15.18}]"} {"utterance_id": "1188-133604-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0015.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After listening to the whole recording, output every instance of an organization name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.19, 1.54, "school"]], "response_mode": "offline", "answer": "[{\"word\":\"school\",\"start\":1.19,\"end\":1.54}]"} {"utterance_id": "4446-2275-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0007.flac", "task_type": "lexical_detection", "query": "pushed", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"pushed\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.82, "pushed"]], "response_mode": "offline", "answer": "[{\"word\":\"pushed\",\"start\":0.53,\"end\":0.82}]"} {"utterance_id": "2961-961-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0019.flac", "task_type": "lexical_detection", "query": "peril", "prompt": "After the audio ends, extract all occurrences of the word \"peril\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.75, 1.36, "peril"]], "response_mode": "offline", "answer": "[{\"word\":\"peril\",\"start\":0.75,\"end\":1.36}]"} {"utterance_id": "1089-134686-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0030.flac", "task_type": "lexical_detection", "query": "mistake", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"mistake\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.9, 2.53, "mistake"]], "response_mode": "offline", "answer": "[{\"word\":\"mistake\",\"start\":1.9,\"end\":2.53}]"} {"utterance_id": "672-122797-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0004.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.66, 5.05, "hare"]], "response_mode": "offline", "answer": "[{\"word\":\"hare\",\"start\":4.66,\"end\":5.05}]"} {"utterance_id": "3570-5695-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0008.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.68, 1.89, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":1.68,\"end\":1.89}]"} {"utterance_id": "1284-1181-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0004.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Process the full audio offline. Find every instance of a country name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.54, 2.99, "oz"]], "response_mode": "offline", "answer": "[{\"word\":\"oz\",\"start\":2.54,\"end\":2.99}]"} {"utterance_id": "5105-28240-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0005.flac", "task_type": "lexical_detection", "query": "extent", "prompt": "After the audio ends, extract all occurrences of the word \"extent\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.1, 3.66, "extent"]], "response_mode": "offline", "answer": "[{\"word\":\"extent\",\"start\":3.1,\"end\":3.66}]"} {"utterance_id": "672-122797-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0008.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.63, 4.35, "fir tree"], [11.02, 11.66, "trees"], [17.78, 18.37, "trees"]], "response_mode": "offline", "answer": "[{\"word\":\"fir tree\",\"start\":3.63,\"end\":4.35},{\"word\":\"trees\",\"start\":11.02,\"end\":11.66},{\"word\":\"trees\",\"start\":17.78,\"end\":18.37}]"} {"utterance_id": "121-121726-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0014.flac", "task_type": "lexical_detection", "query": "horse", "prompt": "When the utterance is complete, report all instances of the word \"horse\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.96, 2.44, "horse"]], "response_mode": "offline", "answer": "[{\"word\":\"horse\",\"start\":1.96,\"end\":2.44}]"} {"utterance_id": "1995-1826-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0005.flac", "task_type": "lexical_detection", "query": "elementary", "prompt": "After the audio ends, extract all occurrences of the word \"elementary\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.44, 4.17, "elementary"]], "response_mode": "offline", "answer": "[{\"word\":\"elementary\",\"start\":3.44,\"end\":4.17}]"} {"utterance_id": "8230-279154-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0020.flac", "task_type": "lexical_detection", "query": "knowledge", "prompt": "Process the full audio offline. Find every instance of the word \"knowledge\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.67, 4.35, "knowledge"]], "response_mode": "offline", "answer": "[{\"word\":\"knowledge\",\"start\":3.67,\"end\":4.35}]"} {"utterance_id": "5105-28240-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0023.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen carefully to the entire clip, then report every instance of an ordinal number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.26, 4.6, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":4.26,\"end\":4.6}]"} {"utterance_id": "1284-1181-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0017.flac", "task_type": "lexical_counting", "query": "wizard", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"wizard\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.43, 0.74, "one"], [8.27, 8.72, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.43,\"end\":0.74},{\"word\":\"two\",\"start\":8.27,\"end\":8.72}]"} {"utterance_id": "3729-6852-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0013.flac", "task_type": "lexical_detection", "query": "honour", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"honour\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.24, 2.81, "honour"]], "response_mode": "offline", "answer": "[{\"word\":\"honour\",\"start\":2.24,\"end\":2.81}]"} {"utterance_id": "8555-292519-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0000.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "You will hear a complete utterance. When it ends, list every instance of a weather term you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.87, 12.46, "rainbow"]], "response_mode": "offline", "answer": "[{\"word\":\"rainbow\",\"start\":11.87,\"end\":12.46}]"} {"utterance_id": "6829-68771-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0010.flac", "task_type": "lexical_detection", "query": "band", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"band\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.97, 1.35, "band"]], "response_mode": "offline", "answer": "[{\"word\":\"band\",\"start\":0.97,\"end\":1.35}]"} {"utterance_id": "4970-29095-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0007.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.54, 5.73, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.54,\"end\":5.73}]"} {"utterance_id": "2961-960-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0016.flac", "task_type": "lexical_detection", "query": "symposium", "prompt": "After the audio ends, extract all occurrences of the word \"symposium\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.56, 7.26, "symposium"]], "response_mode": "offline", "answer": "[{\"word\":\"symposium\",\"start\":6.56,\"end\":7.26}]"} {"utterance_id": "3575-170457-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0051.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Process the full audio offline. Find every instance of a food name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 0.62, "tea"], [2.54, 2.8, "meal"]], "response_mode": "offline", "answer": "[{\"word\":\"tea\",\"start\":0.45,\"end\":0.62},{\"word\":\"meal\",\"start\":2.54,\"end\":2.8}]"} {"utterance_id": "4970-29093-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0013.flac", "task_type": "lexical_detection", "query": "dignified", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"dignified\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.16, 5.69, "dignified"]], "response_mode": "offline", "answer": "[{\"word\":\"dignified\",\"start\":5.16,\"end\":5.69}]"} {"utterance_id": "4507-16021-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0046.flac", "task_type": "lexical_detection", "query": "grief", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"grief\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.28, 2.77, "grief"]], "response_mode": "offline", "answer": "[{\"word\":\"grief\",\"start\":2.28,\"end\":2.77}]"} {"utterance_id": "5639-40744-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0012.flac", "task_type": "lexical_detection", "query": "moonlight", "prompt": "After listening to the whole recording, output every instance of the word \"moonlight\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.6, 3.14, "moonlight"]], "response_mode": "offline", "answer": "[{\"word\":\"moonlight\",\"start\":2.6,\"end\":3.14}]"} {"utterance_id": "7021-85628-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0026.flac", "task_type": "lexical_detection", "query": "son", "prompt": "Process the full audio offline. Find every instance of the word \"son\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.41, 1.79, "son"]], "response_mode": "offline", "answer": "[{\"word\":\"son\",\"start\":1.41,\"end\":1.79}]"} {"utterance_id": "1221-135767-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0023.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.36, 7.88, "reverend mister blackstone"]], "response_mode": "offline", "answer": "[{\"word\":\"reverend mister blackstone\",\"start\":6.36,\"end\":7.88}]"} {"utterance_id": "260-123440-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0017.flac", "task_type": "lexical_detection", "query": "queer", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"queer\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.15, 1.47, "queer"]], "response_mode": "offline", "answer": "[{\"word\":\"queer\",\"start\":1.15,\"end\":1.47}]"} {"utterance_id": "1188-133604-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.23, 2.73, "turner"]], "response_mode": "offline", "answer": "[{\"word\":\"turner\",\"start\":2.23,\"end\":2.73}]"} {"utterance_id": "4992-41806-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0003.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "After the audio ends, extract all occurrences of a time of day. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.4, 8.76, "night"]], "response_mode": "offline", "answer": "[{\"word\":\"night\",\"start\":8.4,\"end\":8.76}]"} {"utterance_id": "5142-36377-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0014.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[6.71, 7.07, "head"], [8.57, 9.13, "shoulders"], [10.19, 10.49, "eyes"], [24.9, 25.31, "ears"]], "response_mode": "offline", "answer": "[{\"word\":\"head\",\"start\":6.71,\"end\":7.07},{\"word\":\"shoulders\",\"start\":8.57,\"end\":9.13},{\"word\":\"eyes\",\"start\":10.19,\"end\":10.49},{\"word\":\"ears\",\"start\":24.9,\"end\":25.31}]"} {"utterance_id": "237-134493-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0001.flac", "task_type": "lexical_detection", "query": "wheat", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"wheat\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.22, 6.5, "wheat"]], "response_mode": "offline", "answer": "[{\"word\":\"wheat\",\"start\":6.22,\"end\":6.5}]"} {"utterance_id": "1188-133604-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0028.flac", "task_type": "semantic_counting", "query": "ordinal", "prompt": "Once the audio has finished, list each occurrence of an ordinal number as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[0.96, 1.57, "one"], [8.86, 9.18, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.96,\"end\":1.57},{\"word\":\"two\",\"start\":8.86,\"end\":9.18}]"} {"utterance_id": "237-134493-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0005.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.19, 6.54, "head"], [7.78, 8.28, "eyes"], [10.13, 10.62, "brow"]], "response_mode": "offline", "answer": "[{\"word\":\"head\",\"start\":6.19,\"end\":6.54},{\"word\":\"eyes\",\"start\":7.78,\"end\":8.28},{\"word\":\"brow\",\"start\":10.13,\"end\":10.62}]"} {"utterance_id": "1188-133604-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0005.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[0.84, 1.03, "head"], [3.02, 3.36, "beak"]], "response_mode": "offline", "answer": "[{\"word\":\"head\",\"start\":0.84,\"end\":1.03},{\"word\":\"beak\",\"start\":3.02,\"end\":3.36}]"} {"utterance_id": "8555-284447-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0012.flac", "task_type": "lexical_detection", "query": "captain", "prompt": "After the audio ends, extract all occurrences of the word \"captain\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.56, 0.99, "captain"]], "response_mode": "offline", "answer": "[{\"word\":\"captain\",\"start\":0.56,\"end\":0.99}]"} {"utterance_id": "2830-3980-0076", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0076.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.68, 1.15, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":0.68,\"end\":1.15}]"} {"utterance_id": "5639-40744-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0022.flac", "task_type": "lexical_detection", "query": "utmost", "prompt": "After the audio ends, extract all occurrences of the word \"utmost\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.15, 5.57, "utmost"]], "response_mode": "offline", "answer": "[{\"word\":\"utmost\",\"start\":5.15,\"end\":5.57}]"} {"utterance_id": "7021-79730-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0000.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.3, 0.7, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":0.3,\"end\":0.7}]"} {"utterance_id": "3729-6852-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0007.flac", "task_type": "lexical_detection", "query": "marivaux", "prompt": "When the utterance is complete, report all instances of the word \"marivaux\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.8, 6.49, "marivaux"]], "response_mode": "offline", "answer": "[{\"word\":\"marivaux\",\"start\":5.8,\"end\":6.49}]"} {"utterance_id": "8455-210777-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0005.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Process the full audio offline. Find every instance of a food name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.7, 5.46, "rose water"]], "response_mode": "offline", "answer": "[{\"word\":\"rose water\",\"start\":4.7,\"end\":5.46}]"} {"utterance_id": "8463-294828-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0010.flac", "task_type": "lexical_detection", "query": "whatever", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"whatever\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.71, 4.31, "whatever"]], "response_mode": "offline", "answer": "[{\"word\":\"whatever\",\"start\":3.71,\"end\":4.31}]"} {"utterance_id": "672-122797-0070", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0070.flac", "task_type": "lexical_detection", "query": "star", "prompt": "After listening to the whole recording, output every instance of the word \"star\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.14, 1.61, "star"]], "response_mode": "offline", "answer": "[{\"word\":\"star\",\"start\":1.14,\"end\":1.61}]"} {"utterance_id": "672-122797-0072", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0072.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "You will hear a complete utterance. When it ends, list every instance of a plant name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.61, 3.1, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":2.61,\"end\":3.1}]"} {"utterance_id": "5142-36377-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0021.flac", "task_type": "lexical_detection", "query": "animated", "prompt": "After listening to the whole recording, output every instance of the word \"animated\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.43, 6.88, "animated"]], "response_mode": "offline", "answer": "[{\"word\":\"animated\",\"start\":6.43,\"end\":6.88}]"} {"utterance_id": "1188-133604-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0041.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.87, 1.34, "gleaner"]], "response_mode": "offline", "answer": "[{\"word\":\"gleaner\",\"start\":0.87,\"end\":1.34}]"} {"utterance_id": "260-123288-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0012.flac", "task_type": "lexical_detection", "query": "never", "prompt": "Process the full audio offline. Find every instance of the word \"never\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.94, 3.5, "never"]], "response_mode": "offline", "answer": "[{\"word\":\"never\",\"start\":2.94,\"end\":3.5}]"} {"utterance_id": "3570-5695-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0008.flac", "task_type": "lexical_detection", "query": "methods", "prompt": "When the utterance is complete, report all instances of the word \"methods\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.89, 2.36, "methods"]], "response_mode": "offline", "answer": "[{\"word\":\"methods\",\"start\":1.89,\"end\":2.36}]"} {"utterance_id": "1995-1826-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0020.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.33, 0.92, "miss taylor"]], "response_mode": "offline", "answer": "[{\"word\":\"miss taylor\",\"start\":0.33,\"end\":0.92}]"} {"utterance_id": "4446-2271-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0022.flac", "task_type": "lexical_detection", "query": "glad", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"glad\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.58, 0.91, "glad"]], "response_mode": "offline", "answer": "[{\"word\":\"glad\",\"start\":0.58,\"end\":0.91}]"} {"utterance_id": "4992-41806-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0001.flac", "task_type": "lexical_detection", "query": "ceremonies", "prompt": "After listening to the whole recording, output every instance of the word \"ceremonies\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.07, 4.76, "ceremonies"]], "response_mode": "offline", "answer": "[{\"word\":\"ceremonies\",\"start\":4.07,\"end\":4.76}]"} {"utterance_id": "8463-294828-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0037.flac", "task_type": "lexical_detection", "query": "burst", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"burst\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.58, 3.94, "burst"]], "response_mode": "offline", "answer": "[{\"word\":\"burst\",\"start\":3.58,\"end\":3.94}]"} {"utterance_id": "7021-85628-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0012.flac", "task_type": "lexical_detection", "query": "eat", "prompt": "Process the full audio offline. Find every instance of the word \"eat\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.04, 1.3, "eat"]], "response_mode": "offline", "answer": "[{\"word\":\"eat\",\"start\":1.04,\"end\":1.3}]"} {"utterance_id": "1188-133604-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0004.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a food name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.9, 4.42, "water"]], "response_mode": "offline", "answer": "[{\"word\":\"water\",\"start\":3.9,\"end\":4.42}]"} {"utterance_id": "1284-1181-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0002.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.59, 0.84, "head"]], "response_mode": "offline", "answer": "[{\"word\":\"head\",\"start\":0.59,\"end\":0.84}]"} {"utterance_id": "1580-141083-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0001.flac", "task_type": "lexical_detection", "query": "occurred", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"occurred\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.22, 9.73, "occurred"]], "response_mode": "offline", "answer": "[{\"word\":\"occurred\",\"start\":9.22,\"end\":9.73}]"} {"utterance_id": "5105-28240-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0004.flac", "task_type": "lexical_detection", "query": "rival", "prompt": "After listening to the whole recording, output every instance of the word \"rival\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.14, 5.61, "rival"]], "response_mode": "offline", "answer": "[{\"word\":\"rival\",\"start\":5.14,\"end\":5.61}]"} {"utterance_id": "2830-3980-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0052.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.22, 1.7, "christ"]], "response_mode": "offline", "answer": "[{\"word\":\"christ\",\"start\":1.22,\"end\":1.7}]"} {"utterance_id": "4446-2275-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0040.flac", "task_type": "lexical_detection", "query": "successful", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"successful\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.85, 3.56, "successful"]], "response_mode": "offline", "answer": "[{\"word\":\"successful\",\"start\":2.85,\"end\":3.56}]"} {"utterance_id": "61-70970-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0002.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.87, 2.32, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":1.87,\"end\":2.32}]"} {"utterance_id": "7176-92135-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0042.flac", "task_type": "lexical_counting", "query": "hero", "prompt": "Process the full audio offline. For every occurrence of the word \"hero\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.87, 1.2, "one"], [4.74, 4.99, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.87,\"end\":1.2},{\"word\":\"two\",\"start\":4.74,\"end\":4.99}]"} {"utterance_id": "3570-5695-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0004.flac", "task_type": "lexical_detection", "query": "achieved", "prompt": "Process the full audio offline. Find every instance of the word \"achieved\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.3, 1.81, "achieved"], [5.57, 6.0, "achieved"]], "response_mode": "offline", "answer": "[{\"word\":\"achieved\",\"start\":1.3,\"end\":1.81},{\"word\":\"achieved\",\"start\":5.57,\"end\":6.0}]"} {"utterance_id": "237-134500-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0023.flac", "task_type": "lexical_detection", "query": "carl", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"carl\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.4, 2.88, "carl"]], "response_mode": "offline", "answer": "[{\"word\":\"carl\",\"start\":2.4,\"end\":2.88}]"} {"utterance_id": "4970-29095-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0030.flac", "task_type": "lexical_detection", "query": "unjust", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"unjust\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.39, 1.94, "unjust"]], "response_mode": "offline", "answer": "[{\"word\":\"unjust\",\"start\":1.39,\"end\":1.94}]"} {"utterance_id": "4446-2275-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0035.flac", "task_type": "lexical_detection", "query": "angrily", "prompt": "Once the audio has finished, identify every instance of the word \"angrily\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.98, 2.48, "angrily"]], "response_mode": "offline", "answer": "[{\"word\":\"angrily\",\"start\":1.98,\"end\":2.48}]"} {"utterance_id": "1221-135767-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0005.flac", "task_type": "semantic_detection", "query": "color", "prompt": "When the utterance is complete, report all instances of a color name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.81, 1.27, "scarlet"], [3.28, 3.74, "scarlet"]], "response_mode": "offline", "answer": "[{\"word\":\"scarlet\",\"start\":0.81,\"end\":1.27},{\"word\":\"scarlet\",\"start\":3.28,\"end\":3.74}]"} {"utterance_id": "61-70970-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0039.flac", "task_type": "lexical_detection", "query": "grave", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"grave\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.99, 2.44, "grave"]], "response_mode": "offline", "answer": "[{\"word\":\"grave\",\"start\":1.99,\"end\":2.44}]"} {"utterance_id": "6829-68771-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0012.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.87, 1.21, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":0.87,\"end\":1.21}]"} {"utterance_id": "908-157963-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0009.flac", "task_type": "lexical_detection", "query": "utter", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"utter\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.9, 3.22, "utter"]], "response_mode": "offline", "answer": "[{\"word\":\"utter\",\"start\":2.9,\"end\":3.22}]"} {"utterance_id": "2300-131720-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0004.flac", "task_type": "lexical_detection", "query": "insistence", "prompt": "After the audio ends, extract all occurrences of the word \"insistence\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.26, 1.93, "insistence"]], "response_mode": "offline", "answer": "[{\"word\":\"insistence\",\"start\":1.26,\"end\":1.93}]"} {"utterance_id": "908-31957-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0002.flac", "task_type": "lexical_counting", "query": "wrong", "prompt": "Listen to the entire clip, then report the count sequence for the word \"wrong\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.92, 1.17, "one"], [3.45, 3.79, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.92,\"end\":1.17},{\"word\":\"two\",\"start\":3.45,\"end\":3.79}]"} {"utterance_id": "6829-68771-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0025.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.37, 7.71, "beth"]], "response_mode": "offline", "answer": "[{\"word\":\"beth\",\"start\":7.37,\"end\":7.71}]"} {"utterance_id": "260-123288-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0001.flac", "task_type": "lexical_detection", "query": "term", "prompt": "When the utterance is complete, report all instances of the word \"term\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.82, 3.22, "term"]], "response_mode": "offline", "answer": "[{\"word\":\"term\",\"start\":2.82,\"end\":3.22}]"} {"utterance_id": "1188-133604-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0007.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a number with word/start/end. If none, reply [].", "labels": [[3.74, 3.98, "one"], [6.15, 6.32, "two"], [9.83, 10.01, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.74,\"end\":3.98},{\"word\":\"two\",\"start\":6.15,\"end\":6.32},{\"word\":\"three\",\"start\":9.83,\"end\":10.01}]"} {"utterance_id": "4077-13751-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0002.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen carefully to the entire clip, then report every instance of an organization name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.35, 7.03, "organization"]], "response_mode": "offline", "answer": "[{\"word\":\"organization\",\"start\":6.35,\"end\":7.03}]"} {"utterance_id": "1320-122617-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0019.flac", "task_type": "lexical_detection", "query": "painful", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"painful\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.94, 7.38, "painful"]], "response_mode": "offline", "answer": "[{\"word\":\"painful\",\"start\":6.94,\"end\":7.38}]"} {"utterance_id": "1580-141083-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0036.flac", "task_type": "lexical_detection", "query": "glare", "prompt": "When the utterance is complete, report all instances of the word \"glare\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.23, 2.55, "glare"]], "response_mode": "offline", "answer": "[{\"word\":\"glare\",\"start\":2.23,\"end\":2.55}]"} {"utterance_id": "6829-68769-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0020.flac", "task_type": "lexical_detection", "query": "pleasant", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"pleasant\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.76, 3.12, "pleasant"]], "response_mode": "offline", "answer": "[{\"word\":\"pleasant\",\"start\":2.76,\"end\":3.12}]"} {"utterance_id": "2094-142345-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0050.flac", "task_type": "lexical_detection", "query": "he'd", "prompt": "After listening to the whole recording, output every instance of the word \"he'd\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.41, 2.59, "he'd"]], "response_mode": "offline", "answer": "[{\"word\":\"he'd\",\"start\":2.41,\"end\":2.59}]"} {"utterance_id": "5142-33396-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0039.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.66, 1.0, "nine"]], "response_mode": "offline", "answer": "[{\"word\":\"nine\",\"start\":0.66,\"end\":1.0}]"} {"utterance_id": "5142-33396-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0022.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.0, 1.6, "olaf"]], "response_mode": "offline", "answer": "[{\"word\":\"olaf\",\"start\":1.0,\"end\":1.6}]"} {"utterance_id": "2961-960-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0020.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[5.11, 5.67, "plato"], [8.63, 9.55, "lucretius"]], "response_mode": "offline", "answer": "[{\"word\":\"plato\",\"start\":5.11,\"end\":5.67},{\"word\":\"lucretius\",\"start\":8.63,\"end\":9.55}]"} {"utterance_id": "5683-32866-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0021.flac", "task_type": "lexical_detection", "query": "mood", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"mood\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.39, 4.83, "mood"]], "response_mode": "offline", "answer": "[{\"word\":\"mood\",\"start\":4.39,\"end\":4.83}]"} {"utterance_id": "121-127105-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0012.flac", "task_type": "lexical_detection", "query": "wasn't", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"wasn't\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.2, 0.52, "wasn't"]], "response_mode": "offline", "answer": "[{\"word\":\"wasn't\",\"start\":0.2,\"end\":0.52}]"} {"utterance_id": "237-134493-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0005.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Process the full audio offline. Find every instance of a color name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.49, 7.78, "gray"]], "response_mode": "offline", "answer": "[{\"word\":\"gray\",\"start\":7.49,\"end\":7.78}]"} {"utterance_id": "5683-32879-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0016.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the full audio. After it finishes, find every instance of a family relation. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.7, 9.19, "cousin's"]], "response_mode": "offline", "answer": "[{\"word\":\"cousin's\",\"start\":8.7,\"end\":9.19}]"} {"utterance_id": "3729-6852-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0009.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "When the utterance is complete, report how a person's name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.51, 5.06, "one"], [16.6, 17.17, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.51,\"end\":5.06},{\"word\":\"two\",\"start\":16.6,\"end\":17.17}]"} {"utterance_id": "237-134500-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0037.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.54, 1.05, "emil"]], "response_mode": "offline", "answer": "[{\"word\":\"emil\",\"start\":0.54,\"end\":1.05}]"} {"utterance_id": "121-127105-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0036.flac", "task_type": "lexical_detection", "query": "reward", "prompt": "Process the full audio offline. Find every instance of the word \"reward\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.42, 2.1, "reward"]], "response_mode": "offline", "answer": "[{\"word\":\"reward\",\"start\":1.42,\"end\":2.1}]"} {"utterance_id": "4507-16021-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0030.flac", "task_type": "lexical_detection", "query": "nation", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"nation\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.39, 3.96, "nation"]], "response_mode": "offline", "answer": "[{\"word\":\"nation\",\"start\":3.39,\"end\":3.96}]"} {"utterance_id": "4446-2275-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0039.flac", "task_type": "lexical_detection", "query": "know", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"know\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.83, 1.05, "know"]], "response_mode": "offline", "answer": "[{\"word\":\"know\",\"start\":0.83,\"end\":1.05}]"} {"utterance_id": "1188-133604-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0027.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Process the full audio offline. Find every instance of a weather term and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.59, 9.01, "mist"]], "response_mode": "offline", "answer": "[{\"word\":\"mist\",\"start\":8.59,\"end\":9.01}]"} {"utterance_id": "260-123440-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0006.flac", "task_type": "lexical_detection", "query": "changed", "prompt": "Process the full audio offline. Find every instance of the word \"changed\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.21, 1.82, "changed"]], "response_mode": "offline", "answer": "[{\"word\":\"changed\",\"start\":1.21,\"end\":1.82}]"} {"utterance_id": "2300-131720-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0015.flac", "task_type": "lexical_counting", "query": "speed", "prompt": "Process the full audio offline. For every occurrence of the word \"speed\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.92, 2.43, "one"], [5.66, 6.26, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.92,\"end\":2.43},{\"word\":\"two\",\"start\":5.66,\"end\":6.26}]"} {"utterance_id": "4970-29095-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0016.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.34, 3.66, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":3.34,\"end\":3.66}]"} {"utterance_id": "237-126133-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0018.flac", "task_type": "lexical_detection", "query": "fault", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"fault\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.15, 3.57, "fault"]], "response_mode": "offline", "answer": "[{\"word\":\"fault\",\"start\":3.15,\"end\":3.57}]"} {"utterance_id": "237-134500-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0023.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.4, 2.88, "carl"]], "response_mode": "offline", "answer": "[{\"word\":\"carl\",\"start\":2.4,\"end\":2.88}]"} {"utterance_id": "3729-6852-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0009.flac", "task_type": "lexical_counting", "query": "silvia", "prompt": "After listening to the whole recording, output timed count words for the word \"silvia\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[4.51, 5.06, "one"], [16.6, 17.17, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.51,\"end\":5.06},{\"word\":\"two\",\"start\":16.6,\"end\":17.17}]"} {"utterance_id": "4077-13751-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0015.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Once the audio has finished, identify every instance of a family relation. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.66, 3.21, "families"]], "response_mode": "offline", "answer": "[{\"word\":\"families\",\"start\":2.66,\"end\":3.21}]"} {"utterance_id": "1221-135767-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0024.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.37, 2.1, "rose bushes"], [3.5, 3.92, "rose"]], "response_mode": "offline", "answer": "[{\"word\":\"rose bushes\",\"start\":1.37,\"end\":2.1},{\"word\":\"rose\",\"start\":3.5,\"end\":3.92}]"} {"utterance_id": "3570-5694-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0018.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.42, 3.61, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":3.42,\"end\":3.61}]"} {"utterance_id": "7021-79740-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.61, 2.1, "jane"], [6.21, 6.65, "jane"]], "response_mode": "offline", "answer": "[{\"word\":\"jane\",\"start\":1.61,\"end\":2.1},{\"word\":\"jane\",\"start\":6.21,\"end\":6.65}]"} {"utterance_id": "5639-40744-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0033.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.77, 4.18, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":3.77,\"end\":4.18}]"} {"utterance_id": "4970-29095-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0031.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[9.66, 10.17, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":9.66,\"end\":10.17}]"} {"utterance_id": "8463-294828-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0038.flac", "task_type": "semantic_detection", "query": "city", "prompt": "You will hear a complete utterance. When it ends, list every instance of a city name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.84, 12.7, "new york city"]], "response_mode": "offline", "answer": "[{\"word\":\"new york city\",\"start\":11.84,\"end\":12.7}]"} {"utterance_id": "1089-134686-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0005.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a color name with fields word, start, end (seconds). If none, reply [].", "labels": [[7.91, 8.33, "pale"]], "response_mode": "offline", "answer": "[{\"word\":\"pale\",\"start\":7.91,\"end\":8.33}]"} {"utterance_id": "8463-294828-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0027.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Do not answer until the audio ends. Then list each instance of a city name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.75, 5.18, "paris"]], "response_mode": "offline", "answer": "[{\"word\":\"paris\",\"start\":4.75,\"end\":5.18}]"} {"utterance_id": "3729-6852-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0020.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "When the utterance is complete, report all instances of an organization name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.68, 8.25, "court"]], "response_mode": "offline", "answer": "[{\"word\":\"court\",\"start\":7.68,\"end\":8.25}]"} {"utterance_id": "908-157963-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0019.flac", "task_type": "lexical_detection", "query": "tenfold", "prompt": "Process the full audio offline. Find every instance of the word \"tenfold\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.98, 1.71, "tenfold"]], "response_mode": "offline", "answer": "[{\"word\":\"tenfold\",\"start\":0.98,\"end\":1.71}]"} {"utterance_id": "5142-36377-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0008.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[7.06, 7.58, "sons"]], "response_mode": "offline", "answer": "[{\"word\":\"sons\",\"start\":7.06,\"end\":7.58}]"} {"utterance_id": "5639-40744-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0023.flac", "task_type": "semantic_counting", "query": "kinship", "prompt": "After the audio ends, tally every instance of a family relation. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[8.26, 8.59, "one"], [9.71, 10.1, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":8.26,\"end\":8.59},{\"word\":\"two\",\"start\":9.71,\"end\":10.1}]"} {"utterance_id": "5142-36377-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0000.flac", "task_type": "lexical_detection", "query": "elder", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"elder\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.86, 3.25, "elder"]], "response_mode": "offline", "answer": "[{\"word\":\"elder\",\"start\":2.86,\"end\":3.25}]"} {"utterance_id": "908-31957-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0022.flac", "task_type": "lexical_detection", "query": "leaves", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"leaves\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.5, 11.14, "leaves"]], "response_mode": "offline", "answer": "[{\"word\":\"leaves\",\"start\":10.5,\"end\":11.14}]"} {"utterance_id": "5683-32879-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.68, 8.12, "redman's"]], "response_mode": "offline", "answer": "[{\"word\":\"redman's\",\"start\":7.68,\"end\":8.12}]"} {"utterance_id": "7176-92135-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0023.flac", "task_type": "lexical_detection", "query": "horatio", "prompt": "After listening to the whole recording, output every instance of the word \"horatio\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.57, 5.17, "horatio"]], "response_mode": "offline", "answer": "[{\"word\":\"horatio\",\"start\":4.57,\"end\":5.17}]"} {"utterance_id": "2830-3980-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0047.flac", "task_type": "lexical_detection", "query": "lose", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"lose\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.28, 1.55, "lose"]], "response_mode": "offline", "answer": "[{\"word\":\"lose\",\"start\":1.28,\"end\":1.55}]"} {"utterance_id": "3570-5694-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0010.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen carefully to the entire clip, then report every instance of a money or currency term as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.42, 3.11, "expenditure"]], "response_mode": "offline", "answer": "[{\"word\":\"expenditure\",\"start\":2.42,\"end\":3.11}]"} {"utterance_id": "3570-5695-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0014.flac", "task_type": "semantic_detection", "query": "money", "prompt": "After the audio ends, extract all occurrences of a money or currency term. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.33, 3.78, "savings"], [9.12, 9.79, "pecuniary"]], "response_mode": "offline", "answer": "[{\"word\":\"savings\",\"start\":3.33,\"end\":3.78},{\"word\":\"pecuniary\",\"start\":9.12,\"end\":9.79}]"} {"utterance_id": "5142-33396-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0048.flac", "task_type": "lexical_detection", "query": "hammer", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"hammer\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.49, 0.87, "hammer"]], "response_mode": "offline", "answer": "[{\"word\":\"hammer\",\"start\":0.49,\"end\":0.87}]"} {"utterance_id": "7176-88083-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0010.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.8, 5.25, "bushes"]], "response_mode": "offline", "answer": "[{\"word\":\"bushes\",\"start\":4.8,\"end\":5.25}]"} {"utterance_id": "3575-170457-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0039.flac", "task_type": "lexical_detection", "query": "anne", "prompt": "Process the full audio offline. Find every instance of the word \"anne\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.48, 2.74, "anne"]], "response_mode": "offline", "answer": "[{\"word\":\"anne\",\"start\":2.48,\"end\":2.74}]"} {"utterance_id": "1089-134686-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0006.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[9.59, 10.09, "tail"]], "response_mode": "offline", "answer": "[{\"word\":\"tail\",\"start\":9.59,\"end\":10.09}]"} {"utterance_id": "4077-13754-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0002.flac", "task_type": "lexical_detection", "query": "ordered", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"ordered\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.37, 2.64, "ordered"]], "response_mode": "offline", "answer": "[{\"word\":\"ordered\",\"start\":2.37,\"end\":2.64}]"} {"utterance_id": "6829-68769-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0052.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.48, 2.8, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":2.48,\"end\":2.8}]"} {"utterance_id": "672-122797-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0016.flac", "task_type": "lexical_detection", "query": "grander", "prompt": "When the utterance is complete, report all instances of the word \"grander\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.79, 4.38, "grander"]], "response_mode": "offline", "answer": "[{\"word\":\"grander\",\"start\":3.79,\"end\":4.38}]"} {"utterance_id": "237-126133-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0013.flac", "task_type": "lexical_detection", "query": "thank", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"thank\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.71, 6.04, "thank"]], "response_mode": "offline", "answer": "[{\"word\":\"thank\",\"start\":5.71,\"end\":6.04}]"} {"utterance_id": "4992-23283-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0012.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.47, 3.76, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":3.47,\"end\":3.76}]"} {"utterance_id": "1284-134647-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0001.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen to the full audio. After it finishes, find every instance of a city name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.09, 1.64, "milan"]], "response_mode": "offline", "answer": "[{\"word\":\"milan\",\"start\":1.09,\"end\":1.64}]"} {"utterance_id": "2094-142345-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0004.flac", "task_type": "lexical_detection", "query": "window", "prompt": "Once the audio has finished, identify every instance of the word \"window\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.69, 2.12, "window"]], "response_mode": "offline", "answer": "[{\"word\":\"window\",\"start\":1.69,\"end\":2.12}]"} {"utterance_id": "260-123288-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0015.flac", "task_type": "semantic_detection", "query": "color", "prompt": "You will hear a complete utterance. When it ends, list every instance of a color name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[19.98, 20.3, "white"]], "response_mode": "offline", "answer": "[{\"word\":\"white\",\"start\":19.98,\"end\":20.3}]"} {"utterance_id": "7021-85628-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0023.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.25, 4.64, "king's"]], "response_mode": "offline", "answer": "[{\"word\":\"king's\",\"start\":4.25,\"end\":4.64}]"} {"utterance_id": "1188-133604-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0005.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "When the utterance is complete, report all instances of an animal name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.21, 1.67, "parrot"]], "response_mode": "offline", "answer": "[{\"word\":\"parrot\",\"start\":1.21,\"end\":1.67}]"} {"utterance_id": "1089-134686-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0002.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Do not answer until the audio ends. Then list each instance of a time of day as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.04, 1.77, "nightfall"]], "response_mode": "offline", "answer": "[{\"word\":\"nightfall\",\"start\":1.04,\"end\":1.77}]"} {"utterance_id": "4507-16021-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0016.flac", "task_type": "lexical_detection", "query": "imagine", "prompt": "Once the audio has finished, identify every instance of the word \"imagine\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.05, 1.69, "imagine"]], "response_mode": "offline", "answer": "[{\"word\":\"imagine\",\"start\":1.05,\"end\":1.69}]"} {"utterance_id": "4507-16021-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0042.flac", "task_type": "lexical_detection", "query": "blacker", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"blacker\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.34, 3.84, "blacker"]], "response_mode": "offline", "answer": "[{\"word\":\"blacker\",\"start\":3.34,\"end\":3.84}]"} {"utterance_id": "1284-1181-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0017.flac", "task_type": "lexical_detection", "query": "wizard", "prompt": "After the audio ends, extract all occurrences of the word \"wizard\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.43, 0.74, "wizard"], [8.27, 8.72, "wizard"]], "response_mode": "offline", "answer": "[{\"word\":\"wizard\",\"start\":0.43,\"end\":0.74},{\"word\":\"wizard\",\"start\":8.27,\"end\":8.72}]"} {"utterance_id": "3575-170457-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0001.flac", "task_type": "lexical_detection", "query": "denied", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"denied\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.04, 1.42, "denied"]], "response_mode": "offline", "answer": "[{\"word\":\"denied\",\"start\":1.04,\"end\":1.42}]"} {"utterance_id": "672-122797-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0000.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the full audio. After it finishes, find every instance of a plant name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.77, 3.55, "fir tree"]], "response_mode": "offline", "answer": "[{\"word\":\"fir tree\",\"start\":2.77,\"end\":3.55}]"} {"utterance_id": "1089-134686-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0029.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "After listening to the whole recording, output every instance of a time of day with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.97, 3.58, "afternoon"]], "response_mode": "offline", "answer": "[{\"word\":\"afternoon\",\"start\":2.97,\"end\":3.58}]"} {"utterance_id": "6829-68769-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0033.flac", "task_type": "lexical_detection", "query": "girl", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"girl\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.72, 2.01, "girl"]], "response_mode": "offline", "answer": "[{\"word\":\"girl\",\"start\":1.72,\"end\":2.01}]"} {"utterance_id": "8463-294825-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0016.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[1.61, 3.24, "one twenty eight thousand"]], "response_mode": "offline", "answer": "[{\"word\":\"one twenty eight thousand\",\"start\":1.61,\"end\":3.24}]"} {"utterance_id": "1995-1837-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0001.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.5, 0.86, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":0.5,\"end\":0.86}]"} {"utterance_id": "4992-41806-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0014.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Process the full audio offline. Find every instance of a time of day and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.46, 10.76, "night"], [13.41, 13.91, "evenin's"]], "response_mode": "offline", "answer": "[{\"word\":\"night\",\"start\":10.46,\"end\":10.76},{\"word\":\"evenin's\",\"start\":13.41,\"end\":13.91}]"} {"utterance_id": "4077-13751-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0020.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Process the full audio offline. Find every instance of an organization name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.29, 7.3, "quincy argus"]], "response_mode": "offline", "answer": "[{\"word\":\"quincy argus\",\"start\":6.29,\"end\":7.3}]"} {"utterance_id": "121-121726-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0011.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.33, 2.72, "next"]], "response_mode": "offline", "answer": "[{\"word\":\"next\",\"start\":2.33,\"end\":2.72}]"} {"utterance_id": "121-123852-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123852/121-123852-0000.flac", "task_type": "lexical_counting", "query": "thy", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"thy\" with word/start/end. If none, reply [].", "labels": [[6.89, 7.11, "one"], [8.75, 9.1, "two"], [10.45, 10.68, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.89,\"end\":7.11},{\"word\":\"two\",\"start\":8.75,\"end\":9.1},{\"word\":\"three\",\"start\":10.45,\"end\":10.68}]"} {"utterance_id": "1188-133604-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0000.flac", "task_type": "lexical_detection", "query": "continually", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"continually\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.13, 1.85, "continually"]], "response_mode": "offline", "answer": "[{\"word\":\"continually\",\"start\":1.13,\"end\":1.85}]"} {"utterance_id": "1221-135767-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0002.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.6, 14.42, "legislators"]], "response_mode": "offline", "answer": "[{\"word\":\"legislators\",\"start\":13.6,\"end\":14.42}]"} {"utterance_id": "1995-1837-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0022.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.26, 1.63, "zora"]], "response_mode": "offline", "answer": "[{\"word\":\"zora\",\"start\":1.26,\"end\":1.63}]"} {"utterance_id": "4992-41797-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0019.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.01, 2.38, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":2.01,\"end\":2.38}]"} {"utterance_id": "8230-279154-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0028.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a time of day with fields word, start, end (seconds). If none, reply [].", "labels": [[6.86, 7.34, "morning"]], "response_mode": "offline", "answer": "[{\"word\":\"morning\",\"start\":6.86,\"end\":7.34}]"} {"utterance_id": "672-122797-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0034.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.45, 1.97, "children"]], "response_mode": "offline", "answer": "[{\"word\":\"children\",\"start\":1.45,\"end\":1.97}]"} {"utterance_id": "7021-79730-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0008.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.55, 0.97, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":0.55,\"end\":0.97}]"} {"utterance_id": "4507-16021-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0023.flac", "task_type": "semantic_detection", "query": "food", "prompt": "After listening to the whole recording, output every instance of a food name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.57, 0.91, "sugar"], [3.0, 3.57, "loaf"]], "response_mode": "offline", "answer": "[{\"word\":\"sugar\",\"start\":0.57,\"end\":0.91},{\"word\":\"loaf\",\"start\":3.0,\"end\":3.57}]"} {"utterance_id": "4446-2273-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0035.flac", "task_type": "lexical_detection", "query": "leaned", "prompt": "Process the full audio offline. Find every instance of the word \"leaned\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.63, 0.83, "leaned"]], "response_mode": "offline", "answer": "[{\"word\":\"leaned\",\"start\":0.63,\"end\":0.83}]"} {"utterance_id": "1580-141084-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0023.flac", "task_type": "semantic_detection", "query": "money", "prompt": "After listening to the whole recording, output every instance of a money or currency term with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.75, 8.39, "scholarship"]], "response_mode": "offline", "answer": "[{\"word\":\"scholarship\",\"start\":7.75,\"end\":8.39}]"} {"utterance_id": "2830-3980-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0033.flac", "task_type": "lexical_detection", "query": "galatians", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"galatians\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.04, 1.52, "galatians"]], "response_mode": "offline", "answer": "[{\"word\":\"galatians\",\"start\":1.04,\"end\":1.52}]"} {"utterance_id": "5142-33396-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0044.flac", "task_type": "lexical_detection", "query": "farmer", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"farmer\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.23, 4.78, "farmer"]], "response_mode": "offline", "answer": "[{\"word\":\"farmer\",\"start\":4.23,\"end\":4.78}]"} {"utterance_id": "3729-6852-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0021.flac", "task_type": "lexical_detection", "query": "leave", "prompt": "After the audio ends, extract all occurrences of the word \"leave\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.05, 2.51, "leave"]], "response_mode": "offline", "answer": "[{\"word\":\"leave\",\"start\":2.05,\"end\":2.51}]"} {"utterance_id": "1284-1181-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0009.flac", "task_type": "lexical_detection", "query": "lift", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"lift\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.25, 3.51, "lift"]], "response_mode": "offline", "answer": "[{\"word\":\"lift\",\"start\":3.25,\"end\":3.51}]"} {"utterance_id": "1221-135767-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0006.flac", "task_type": "lexical_detection", "query": "brain", "prompt": "After the audio ends, extract all occurrences of the word \"brain\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.05, 5.58, "brain"]], "response_mode": "offline", "answer": "[{\"word\":\"brain\",\"start\":5.05,\"end\":5.58}]"} {"utterance_id": "3570-5696-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0000.flac", "task_type": "lexical_detection", "query": "expect", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"expect\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.63, 4.03, "expect"]], "response_mode": "offline", "answer": "[{\"word\":\"expect\",\"start\":3.63,\"end\":4.03}]"} {"utterance_id": "4507-16021-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0026.flac", "task_type": "lexical_counting", "query": "civilization", "prompt": "After the audio ends, tally every instance of the word \"civilization\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[18.43, 19.59, "one"], [29.13, 30.43, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":18.43,\"end\":19.59},{\"word\":\"two\",\"start\":29.13,\"end\":30.43}]"} {"utterance_id": "1320-122617-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0022.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.97, 1.33, "children"]], "response_mode": "offline", "answer": "[{\"word\":\"children\",\"start\":0.97,\"end\":1.33}]"} {"utterance_id": "260-123288-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0019.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Listen carefully to the entire clip, then report every instance of a weather term as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.69, 2.12, "storm"]], "response_mode": "offline", "answer": "[{\"word\":\"storm\",\"start\":1.69,\"end\":2.12}]"} {"utterance_id": "8224-274381-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0017.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[0.53, 1.4, "historians"]], "response_mode": "offline", "answer": "[{\"word\":\"historians\",\"start\":0.53,\"end\":1.4}]"} {"utterance_id": "1580-141083-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0031.flac", "task_type": "lexical_detection", "query": "letters", "prompt": "After listening to the whole recording, output every instance of the word \"letters\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.02, 2.43, "letters"]], "response_mode": "offline", "answer": "[{\"word\":\"letters\",\"start\":2.02,\"end\":2.43}]"} {"utterance_id": "672-122797-0067", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0067.flac", "task_type": "lexical_detection", "query": "daylight", "prompt": "After listening to the whole recording, output every instance of the word \"daylight\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.38, 11.88, "daylight"]], "response_mode": "offline", "answer": "[{\"word\":\"daylight\",\"start\":11.38,\"end\":11.88}]"} {"utterance_id": "3570-5696-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0000.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen to the full audio. After it finishes, find every instance of a money or currency term. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.9, 8.57, "pecuniary"]], "response_mode": "offline", "answer": "[{\"word\":\"pecuniary\",\"start\":7.9,\"end\":8.57}]"} {"utterance_id": "121-121726-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0010.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the full audio. After it finishes, find every instance of an organization name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.33, 7.09, "government"]], "response_mode": "offline", "answer": "[{\"word\":\"government\",\"start\":6.33,\"end\":7.09}]"} {"utterance_id": "2830-3980-0055", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0055.flac", "task_type": "lexical_detection", "query": "province", "prompt": "Process the full audio offline. Find every instance of the word \"province\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.46, 2.94, "province"]], "response_mode": "offline", "answer": "[{\"word\":\"province\",\"start\":2.46,\"end\":2.94}]"} {"utterance_id": "1320-122617-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0021.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.59, 3.95, "singer"]], "response_mode": "offline", "answer": "[{\"word\":\"singer\",\"start\":3.59,\"end\":3.95}]"} {"utterance_id": "2094-142345-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0002.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.54, 15.01, "pair"]], "response_mode": "offline", "answer": "[{\"word\":\"pair\",\"start\":14.54,\"end\":15.01}]"} {"utterance_id": "5639-40744-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0023.flac", "task_type": "lexical_counting", "query": "blessings", "prompt": "Process the full audio offline. For every occurrence of the word \"blessings\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.4, 3.03, "one"], [6.05, 6.63, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.4,\"end\":3.03},{\"word\":\"two\",\"start\":6.05,\"end\":6.63}]"} {"utterance_id": "1995-1837-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0020.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.5, 3.08, "ten"]], "response_mode": "offline", "answer": "[{\"word\":\"ten\",\"start\":2.5,\"end\":3.08}]"} {"utterance_id": "1188-133604-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0034.flac", "task_type": "lexical_counting", "query": "order", "prompt": "Process the full audio offline. For every occurrence of the word \"order\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.05, 1.53, "one"], [19.93, 20.47, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.05,\"end\":1.53},{\"word\":\"two\",\"start\":19.93,\"end\":20.47}]"} {"utterance_id": "237-126133-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0017.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a person's name (word, start, end). If none, reply [].", "labels": [[0.8, 1.13, "one"], [2.46, 3.01, "two"], [4.79, 5.27, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.8,\"end\":1.13},{\"word\":\"two\",\"start\":2.46,\"end\":3.01},{\"word\":\"three\",\"start\":4.79,\"end\":5.27}]"} {"utterance_id": "260-123288-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0011.flac", "task_type": "lexical_detection", "query": "loose", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"loose\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.3, 4.79, "loose"]], "response_mode": "offline", "answer": "[{\"word\":\"loose\",\"start\":4.3,\"end\":4.79}]"} {"utterance_id": "260-123286-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0029.flac", "task_type": "lexical_detection", "query": "huge", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"huge\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.61, 1.04, "huge"]], "response_mode": "offline", "answer": "[{\"word\":\"huge\",\"start\":0.61,\"end\":1.04}]"} {"utterance_id": "7021-79730-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0006.flac", "task_type": "semantic_counting", "query": "kinship", "prompt": "After the audio ends, tally every instance of a family relation. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.68, 1.14, "one"], [6.95, 7.5, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.68,\"end\":1.14},{\"word\":\"two\",\"start\":6.95,\"end\":7.5}]"} {"utterance_id": "8230-279154-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0038.flac", "task_type": "lexical_counting", "query": "recognition", "prompt": "After the audio ends, tally every instance of the word \"recognition\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[10.06, 10.91, "one"], [18.55, 19.39, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":10.06,\"end\":10.91},{\"word\":\"two\",\"start\":18.55,\"end\":19.39}]"} {"utterance_id": "7729-102255-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0029.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.83, "ten"]], "response_mode": "offline", "answer": "[{\"word\":\"ten\",\"start\":0.53,\"end\":0.83}]"} {"utterance_id": "121-123852-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123852/121-123852-0003.flac", "task_type": "lexical_detection", "query": "wrought", "prompt": "After listening to the whole recording, output every instance of the word \"wrought\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.36, 9.84, "wrought"]], "response_mode": "offline", "answer": "[{\"word\":\"wrought\",\"start\":9.36,\"end\":9.84}]"} {"utterance_id": "5142-36377-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0000.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.36, 0.52, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.36,\"end\":0.52}]"} {"utterance_id": "1320-122612-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.3, 1.76, "uncas"]], "response_mode": "offline", "answer": "[{\"word\":\"uncas\",\"start\":1.3,\"end\":1.76}]"} {"utterance_id": "1995-1837-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0019.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.56, 5.23, "zora"]], "response_mode": "offline", "answer": "[{\"word\":\"zora\",\"start\":4.56,\"end\":5.23}]"} {"utterance_id": "237-126133-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0003.flac", "task_type": "lexical_detection", "query": "strongest", "prompt": "After listening to the whole recording, output every instance of the word \"strongest\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.95, 3.55, "strongest"]], "response_mode": "offline", "answer": "[{\"word\":\"strongest\",\"start\":2.95,\"end\":3.55}]"} {"utterance_id": "5142-33396-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0046.flac", "task_type": "lexical_detection", "query": "joke", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"joke\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.62, 2.9, "joke"]], "response_mode": "offline", "answer": "[{\"word\":\"joke\",\"start\":2.62,\"end\":2.9}]"} {"utterance_id": "5142-33396-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0046.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.5, 0.82, "beard"]], "response_mode": "offline", "answer": "[{\"word\":\"beard\",\"start\":0.5,\"end\":0.82}]"} {"utterance_id": "7176-88083-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0002.flac", "task_type": "lexical_counting", "query": "red", "prompt": "After listening to the whole recording, output timed count words for the word \"red\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[0.96, 1.51, "one"], [6.86, 7.24, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.96,\"end\":1.51},{\"word\":\"two\",\"start\":6.86,\"end\":7.24}]"} {"utterance_id": "7176-92135-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0028.flac", "task_type": "lexical_detection", "query": "third", "prompt": "When the utterance is complete, report all instances of the word \"third\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.23, 2.71, "third"]], "response_mode": "offline", "answer": "[{\"word\":\"third\",\"start\":2.23,\"end\":2.71}]"} {"utterance_id": "4970-29095-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0036.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a family relation with fields word, start, end (seconds). If none, reply [].", "labels": [[1.35, 1.75, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":1.35,\"end\":1.75}]"} {"utterance_id": "2830-3980-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0026.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.87, 1.27, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":0.87,\"end\":1.27}]"} {"utterance_id": "6930-81414-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0013.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.46, 7.04, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":6.46,\"end\":7.04}]"} {"utterance_id": "2094-142345-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0025.flac", "task_type": "lexical_detection", "query": "sweet", "prompt": "Once the audio has finished, identify every instance of the word \"sweet\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.53, 2.84, "sweet"]], "response_mode": "offline", "answer": "[{\"word\":\"sweet\",\"start\":2.53,\"end\":2.84}]"} {"utterance_id": "7176-92135-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0037.flac", "task_type": "lexical_detection", "query": "introduce", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"introduce\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.08, 1.5, "introduce"]], "response_mode": "offline", "answer": "[{\"word\":\"introduce\",\"start\":1.08,\"end\":1.5}]"} {"utterance_id": "8230-279154-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0036.flac", "task_type": "lexical_detection", "query": "further", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"further\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.65, 1.1, "further"]], "response_mode": "offline", "answer": "[{\"word\":\"further\",\"start\":0.65,\"end\":1.1}]"} {"utterance_id": "4507-16021-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0012.flac", "task_type": "lexical_detection", "query": "why", "prompt": "After listening to the whole recording, output every instance of the word \"why\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.28, 0.77, "why"]], "response_mode": "offline", "answer": "[{\"word\":\"why\",\"start\":0.28,\"end\":0.77}]"} {"utterance_id": "6930-76324-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0028.flac", "task_type": "lexical_detection", "query": "boarded", "prompt": "Process the full audio offline. Find every instance of the word \"boarded\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.68, 2.13, "boarded"]], "response_mode": "offline", "answer": "[{\"word\":\"boarded\",\"start\":1.68,\"end\":2.13}]"} {"utterance_id": "8455-210777-0064", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0064.flac", "task_type": "lexical_detection", "query": "government", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"government\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.06, 3.48, "government"]], "response_mode": "offline", "answer": "[{\"word\":\"government\",\"start\":3.06,\"end\":3.48}]"} {"utterance_id": "8555-284447-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0002.flac", "task_type": "semantic_counting", "query": "food", "prompt": "When the utterance is complete, report how a food name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.37, 1.68, "one"], [7.14, 7.81, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.37,\"end\":1.68},{\"word\":\"two\",\"start\":7.14,\"end\":7.81}]"} {"utterance_id": "8224-274384-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0004.flac", "task_type": "lexical_detection", "query": "bringing", "prompt": "After listening to the whole recording, output every instance of the word \"bringing\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[18.35, 18.78, "bringing"]], "response_mode": "offline", "answer": "[{\"word\":\"bringing\",\"start\":18.35,\"end\":18.78}]"} {"utterance_id": "6930-76324-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0007.flac", "task_type": "lexical_detection", "query": "say", "prompt": "Process the full audio offline. Find every instance of the word \"say\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.98, 1.44, "say"]], "response_mode": "offline", "answer": "[{\"word\":\"say\",\"start\":0.98,\"end\":1.44}]"} {"utterance_id": "1995-1837-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0009.flac", "task_type": "lexical_detection", "query": "week", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"week\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.07, 2.33, "week"]], "response_mode": "offline", "answer": "[{\"word\":\"week\",\"start\":2.07,\"end\":2.33}]"} {"utterance_id": "1320-122617-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0033.flac", "task_type": "lexical_detection", "query": "depend", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"depend\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.47, 4.96, "depend"]], "response_mode": "offline", "answer": "[{\"word\":\"depend\",\"start\":4.47,\"end\":4.96}]"} {"utterance_id": "4992-23283-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 1.43, "missus horton"]], "response_mode": "offline", "answer": "[{\"word\":\"missus horton\",\"start\":0.52,\"end\":1.43}]"} {"utterance_id": "1221-135767-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0009.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.16, 2.67, "child"]], "response_mode": "offline", "answer": "[{\"word\":\"child\",\"start\":2.16,\"end\":2.67}]"} {"utterance_id": "61-70970-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0039.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.66, 6.24, "hands"]], "response_mode": "offline", "answer": "[{\"word\":\"hands\",\"start\":5.66,\"end\":6.24}]"} {"utterance_id": "1320-122617-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0034.flac", "task_type": "lexical_detection", "query": "assurance", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"assurance\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.46, 2.94, "assurance"]], "response_mode": "offline", "answer": "[{\"word\":\"assurance\",\"start\":2.46,\"end\":2.94}]"} {"utterance_id": "8455-210777-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0035.flac", "task_type": "lexical_detection", "query": "ferdinando", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"ferdinando\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.29, 4.03, "ferdinando"]], "response_mode": "offline", "answer": "[{\"word\":\"ferdinando\",\"start\":3.29,\"end\":4.03}]"} {"utterance_id": "4077-13751-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0014.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen to the full audio. After it finishes, find every instance of a city name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.55, 12.24, "far west"]], "response_mode": "offline", "answer": "[{\"word\":\"far west\",\"start\":11.55,\"end\":12.24}]"} {"utterance_id": "7127-75947-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0040.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an organization name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.08, 2.54, "court"]], "response_mode": "offline", "answer": "[{\"word\":\"court\",\"start\":2.08,\"end\":2.54}]"} {"utterance_id": "672-122797-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0006.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Once the audio has finished, identify every instance of a plant name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.85, 7.31, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":6.85,\"end\":7.31}]"} {"utterance_id": "5142-33396-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0031.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Once the audio has finished, identify every instance of a food name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.88, 7.16, "supper"]], "response_mode": "offline", "answer": "[{\"word\":\"supper\",\"start\":6.88,\"end\":7.16}]"} {"utterance_id": "6930-75918-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0016.flac", "task_type": "lexical_detection", "query": "repose", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"repose\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.31, 1.87, "repose"]], "response_mode": "offline", "answer": "[{\"word\":\"repose\",\"start\":1.31,\"end\":1.87}]"} {"utterance_id": "7729-102255-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0015.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After the audio ends, extract all occurrences of an organization name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.9, 1.36, "militia"], [8.79, 9.18, "sharps"], [12.46, 13.48, "government troops"]], "response_mode": "offline", "answer": "[{\"word\":\"militia\",\"start\":0.9,\"end\":1.36},{\"word\":\"sharps\",\"start\":8.79,\"end\":9.18},{\"word\":\"government troops\",\"start\":12.46,\"end\":13.48}]"} {"utterance_id": "1284-1180-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0013.flac", "task_type": "lexical_detection", "query": "unlucky", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"unlucky\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.61, 2.4, "unlucky"]], "response_mode": "offline", "answer": "[{\"word\":\"unlucky\",\"start\":1.61,\"end\":2.4}]"} {"utterance_id": "8463-294825-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0012.flac", "task_type": "lexical_detection", "query": "growing", "prompt": "When the utterance is complete, report all instances of the word \"growing\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.53, 4.88, "growing"]], "response_mode": "offline", "answer": "[{\"word\":\"growing\",\"start\":4.53,\"end\":4.88}]"} {"utterance_id": "5683-32879-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0020.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the full audio. After it finishes, find every instance of a family relation. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.12, 2.54, "cousin's"]], "response_mode": "offline", "answer": "[{\"word\":\"cousin's\",\"start\":2.12,\"end\":2.54}]"} {"utterance_id": "121-127105-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0026.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.47, 0.75, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":0.47,\"end\":0.75}]"} {"utterance_id": "8463-294828-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0012.flac", "task_type": "lexical_detection", "query": "admitting", "prompt": "After listening to the whole recording, output every instance of the word \"admitting\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.49, 3.16, "admitting"]], "response_mode": "offline", "answer": "[{\"word\":\"admitting\",\"start\":2.49,\"end\":3.16}]"} {"utterance_id": "4992-41797-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0006.flac", "task_type": "lexical_detection", "query": "shalt", "prompt": "Process the full audio offline. Find every instance of the word \"shalt\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.36, 1.77, "shalt"]], "response_mode": "offline", "answer": "[{\"word\":\"shalt\",\"start\":1.36,\"end\":1.77}]"} {"utterance_id": "1221-135767-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0024.flac", "task_type": "semantic_detection", "query": "color", "prompt": "When the utterance is complete, report all instances of a color name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.25, 3.5, "red"]], "response_mode": "offline", "answer": "[{\"word\":\"red\",\"start\":3.25,\"end\":3.5}]"} {"utterance_id": "4507-16021-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0009.flac", "task_type": "lexical_detection", "query": "gulf", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"gulf\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.4, 4.09, "gulf"]], "response_mode": "offline", "answer": "[{\"word\":\"gulf\",\"start\":3.4,\"end\":4.09}]"} {"utterance_id": "7021-79740-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0002.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.35, 6.89, "children"]], "response_mode": "offline", "answer": "[{\"word\":\"children\",\"start\":6.35,\"end\":6.89}]"} {"utterance_id": "1580-141084-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0021.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.84, 1.19, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":0.84,\"end\":1.19}]"} {"utterance_id": "5683-32866-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0006.flac", "task_type": "lexical_detection", "query": "worse", "prompt": "Process the full audio offline. Find every instance of the word \"worse\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.59, 4.11, "worse"]], "response_mode": "offline", "answer": "[{\"word\":\"worse\",\"start\":3.59,\"end\":4.11}]"} {"utterance_id": "908-31957-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0025.flac", "task_type": "lexical_detection", "query": "smiles", "prompt": "Once the audio has finished, identify every instance of the word \"smiles\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[9.07, 9.87, "smiles"]], "response_mode": "offline", "answer": "[{\"word\":\"smiles\",\"start\":9.07,\"end\":9.87}]"} {"utterance_id": "4507-16021-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0051.flac", "task_type": "lexical_detection", "query": "evidently", "prompt": "When the utterance is complete, report all instances of the word \"evidently\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.33, 2.14, "evidently"]], "response_mode": "offline", "answer": "[{\"word\":\"evidently\",\"start\":1.33,\"end\":2.14}]"} {"utterance_id": "4446-2273-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0000.flac", "task_type": "lexical_detection", "query": "chin", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"chin\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.62, 6.83, "chin"]], "response_mode": "offline", "answer": "[{\"word\":\"chin\",\"start\":6.62,\"end\":6.83}]"} {"utterance_id": "6829-68769-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.93, 2.33, "kenneth"]], "response_mode": "offline", "answer": "[{\"word\":\"kenneth\",\"start\":1.93,\"end\":2.33}]"} {"utterance_id": "1188-133604-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0029.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.7, 1.91, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":1.7,\"end\":1.91}]"} {"utterance_id": "1221-135766-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0001.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.62, 9.07, "bosom"]], "response_mode": "offline", "answer": "[{\"word\":\"bosom\",\"start\":8.62,\"end\":9.07}]"} {"utterance_id": "4446-2275-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.48, 1.06, "alexander"]], "response_mode": "offline", "answer": "[{\"word\":\"alexander\",\"start\":0.48,\"end\":1.06}]"} {"utterance_id": "1089-134686-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0037.flac", "task_type": "lexical_detection", "query": "kindled", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"kindled\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.34, 2.99, "kindled"]], "response_mode": "offline", "answer": "[{\"word\":\"kindled\",\"start\":2.34,\"end\":2.99}]"} {"utterance_id": "1089-134691-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0023.flac", "task_type": "lexical_detection", "query": "sword", "prompt": "Process the full audio offline. Find every instance of the word \"sword\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.99, 2.35, "sword"]], "response_mode": "offline", "answer": "[{\"word\":\"sword\",\"start\":1.99,\"end\":2.35}]"} {"utterance_id": "1580-141083-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0010.flac", "task_type": "lexical_detection", "query": "careful", "prompt": "After listening to the whole recording, output every instance of the word \"careful\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.47, 3.81, "careful"]], "response_mode": "offline", "answer": "[{\"word\":\"careful\",\"start\":3.47,\"end\":3.81}]"} {"utterance_id": "1995-1837-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0006.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "After listening to the whole recording, output every instance of a weather term with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.97, 2.57, "mists"]], "response_mode": "offline", "answer": "[{\"word\":\"mists\",\"start\":1.97,\"end\":2.57}]"} {"utterance_id": "2094-142345-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0036.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the full audio. After it finishes, find every instance of a family relation. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.02, 1.35, "aunt"], [5.75, 6.47, "families"]], "response_mode": "offline", "answer": "[{\"word\":\"aunt\",\"start\":1.02,\"end\":1.35},{\"word\":\"families\",\"start\":5.75,\"end\":6.47}]"} {"utterance_id": "4992-41806-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.35, 1.1, "mother carey"], [2.29, 2.7, "nancy"]], "response_mode": "offline", "answer": "[{\"word\":\"mother carey\",\"start\":0.35,\"end\":1.1},{\"word\":\"nancy\",\"start\":2.29,\"end\":2.7}]"} {"utterance_id": "7021-79730-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0006.flac", "task_type": "lexical_counting", "query": "partly", "prompt": "Listen to the entire clip, then report the count sequence for the word \"partly\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.34, 4.89, "one"], [8.01, 8.58, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.34,\"end\":4.89},{\"word\":\"two\",\"start\":8.01,\"end\":8.58}]"} {"utterance_id": "237-134500-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0008.flac", "task_type": "lexical_detection", "query": "season", "prompt": "Once the audio has finished, identify every instance of the word \"season\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.1, 1.52, "season"]], "response_mode": "offline", "answer": "[{\"word\":\"season\",\"start\":1.1,\"end\":1.52}]"} {"utterance_id": "2830-3980-0066", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0066.flac", "task_type": "lexical_detection", "query": "stands", "prompt": "Once the audio has finished, identify every instance of the word \"stands\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.13, 2.6, "stands"]], "response_mode": "offline", "answer": "[{\"word\":\"stands\",\"start\":2.13,\"end\":2.6}]"} {"utterance_id": "4970-29095-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0022.flac", "task_type": "lexical_detection", "query": "yearly", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"yearly\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.64, 1.95, "yearly"]], "response_mode": "offline", "answer": "[{\"word\":\"yearly\",\"start\":1.64,\"end\":1.95}]"} {"utterance_id": "1221-135767-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0003.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.74, 7.16, "pig"]], "response_mode": "offline", "answer": "[{\"word\":\"pig\",\"start\":6.74,\"end\":7.16}]"} {"utterance_id": "3575-170457-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0053.flac", "task_type": "lexical_detection", "query": "charlotte", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"charlotte\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.2, 3.6, "charlotte"]], "response_mode": "offline", "answer": "[{\"word\":\"charlotte\",\"start\":3.2,\"end\":3.6}]"} {"utterance_id": "7176-88083-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0026.flac", "task_type": "lexical_detection", "query": "wings", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"wings\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.85, 3.18, "wings"]], "response_mode": "offline", "answer": "[{\"word\":\"wings\",\"start\":2.85,\"end\":3.18}]"} {"utterance_id": "7021-79740-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0003.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[3.68, 3.84, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.68,\"end\":3.84}]"} {"utterance_id": "908-31957-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0023.flac", "task_type": "lexical_counting", "query": "thee", "prompt": "When the utterance is complete, report how the word \"thee\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.91, 1.08, "one"], [5.18, 5.35, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.91,\"end\":1.08},{\"word\":\"two\",\"start\":5.18,\"end\":5.35}]"} {"utterance_id": "121-121726-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0007.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "After the audio ends, extract all occurrences of an animal name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 0.99, "horse"]], "response_mode": "offline", "answer": "[{\"word\":\"horse\",\"start\":0.51,\"end\":0.99}]"} {"utterance_id": "8230-279154-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0024.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen carefully to the entire clip, then report every instance of an ordinal number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.76, 1.39, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":0.76,\"end\":1.39}]"} {"utterance_id": "1089-134686-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0015.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen carefully to the entire clip, then report every instance of a time of day as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.38, 0.83, "dusk"]], "response_mode": "offline", "answer": "[{\"word\":\"dusk\",\"start\":0.38,\"end\":0.83}]"} {"utterance_id": "4970-29095-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0000.flac", "task_type": "lexical_detection", "query": "things", "prompt": "After listening to the whole recording, output every instance of the word \"things\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.87, 2.46, "things"]], "response_mode": "offline", "answer": "[{\"word\":\"things\",\"start\":1.87,\"end\":2.46}]"} {"utterance_id": "1089-134686-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0019.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.27, 2.08, "saint james"]], "response_mode": "offline", "answer": "[{\"word\":\"saint james\",\"start\":1.27,\"end\":2.08}]"} {"utterance_id": "3729-6852-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0036.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.62, 0.9, "king"], [4.28, 4.93, "roi"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":0.62,\"end\":0.9},{\"word\":\"roi\",\"start\":4.28,\"end\":4.93}]"} {"utterance_id": "2300-131720-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0027.flac", "task_type": "lexical_detection", "query": "measured", "prompt": "When the utterance is complete, report all instances of the word \"measured\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.36, 3.89, "measured"]], "response_mode": "offline", "answer": "[{\"word\":\"measured\",\"start\":3.36,\"end\":3.89}]"} {"utterance_id": "1320-122617-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0024.flac", "task_type": "lexical_detection", "query": "note", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"note\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.12, 3.36, "note"]], "response_mode": "offline", "answer": "[{\"word\":\"note\",\"start\":3.12,\"end\":3.36}]"} {"utterance_id": "4970-29095-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0031.flac", "task_type": "lexical_counting", "query": "exactly", "prompt": "After listening to the whole recording, output timed count words for the word \"exactly\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[1.27, 2.12, "one"], [12.45, 13.0, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.27,\"end\":2.12},{\"word\":\"two\",\"start\":12.45,\"end\":13.0}]"} {"utterance_id": "4970-29095-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0006.flac", "task_type": "lexical_detection", "query": "willing", "prompt": "When the utterance is complete, report all instances of the word \"willing\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.18, 1.57, "willing"]], "response_mode": "offline", "answer": "[{\"word\":\"willing\",\"start\":1.18,\"end\":1.57}]"} {"utterance_id": "237-134493-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0001.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen carefully to the entire clip, then report every instance of a color name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.9, 3.22, "white"]], "response_mode": "offline", "answer": "[{\"word\":\"white\",\"start\":2.9,\"end\":3.22}]"} {"utterance_id": "3575-170457-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0046.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen to the full audio. After it finishes, find every instance of a city name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.74, 6.55, "haworth"]], "response_mode": "offline", "answer": "[{\"word\":\"haworth\",\"start\":5.74,\"end\":6.55}]"} {"utterance_id": "1580-141083-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0015.flac", "task_type": "lexical_detection", "query": "proofs", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"proofs\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.71, 2.1, "proofs"]], "response_mode": "offline", "answer": "[{\"word\":\"proofs\",\"start\":1.71,\"end\":2.1}]"} {"utterance_id": "6930-76324-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0008.flac", "task_type": "lexical_detection", "query": "saw", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"saw\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.07, 2.25, "saw"]], "response_mode": "offline", "answer": "[{\"word\":\"saw\",\"start\":2.07,\"end\":2.25}]"} {"utterance_id": "7176-92135-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0033.flac", "task_type": "lexical_detection", "query": "silence", "prompt": "After the audio ends, extract all occurrences of the word \"silence\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.9, 1.27, "silence"]], "response_mode": "offline", "answer": "[{\"word\":\"silence\",\"start\":0.9,\"end\":1.27}]"} {"utterance_id": "4970-29093-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0022.flac", "task_type": "lexical_detection", "query": "uncle", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"uncle\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.06, 1.41, "uncle"]], "response_mode": "offline", "answer": "[{\"word\":\"uncle\",\"start\":1.06,\"end\":1.41}]"} {"utterance_id": "6829-68771-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0022.flac", "task_type": "lexical_detection", "query": "attend", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"attend\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 0.83, "attend"]], "response_mode": "offline", "answer": "[{\"word\":\"attend\",\"start\":0.45,\"end\":0.83}]"} {"utterance_id": "1995-1826-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0024.flac", "task_type": "semantic_counting", "query": "color", "prompt": "Listen to the full audio. Count every occurrence of a color name. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.34, 0.87, "one"], [2.42, 2.85, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.34,\"end\":0.87},{\"word\":\"two\",\"start\":2.42,\"end\":2.85}]"} {"utterance_id": "5142-33396-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.04, 1.47, "harald"]], "response_mode": "offline", "answer": "[{\"word\":\"harald\",\"start\":1.04,\"end\":1.47}]"} {"utterance_id": "4970-29095-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0003.flac", "task_type": "lexical_counting", "query": "thee", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"thee\" with word/start/end. If none, reply [].", "labels": [[1.05, 1.21, "one"], [6.26, 6.44, "two"], [7.56, 7.85, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.05,\"end\":1.21},{\"word\":\"two\",\"start\":6.26,\"end\":6.44},{\"word\":\"three\",\"start\":7.56,\"end\":7.85}]"} {"utterance_id": "8224-274381-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0003.flac", "task_type": "lexical_detection", "query": "proposed", "prompt": "After the audio ends, extract all occurrences of the word \"proposed\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.74, 5.36, "proposed"]], "response_mode": "offline", "answer": "[{\"word\":\"proposed\",\"start\":4.74,\"end\":5.36}]"} {"utterance_id": "2961-960-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0022.flac", "task_type": "lexical_detection", "query": "enabled", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"enabled\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.11, 4.68, "enabled"]], "response_mode": "offline", "answer": "[{\"word\":\"enabled\",\"start\":4.11,\"end\":4.68}]"} {"utterance_id": "7021-79740-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0003.flac", "task_type": "lexical_detection", "query": "idea", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"idea\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.75, 1.32, "idea"]], "response_mode": "offline", "answer": "[{\"word\":\"idea\",\"start\":0.75,\"end\":1.32}]"} {"utterance_id": "4077-13754-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0016.flac", "task_type": "lexical_detection", "query": "runs", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"runs\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.82, 2.12, "runs"]], "response_mode": "offline", "answer": "[{\"word\":\"runs\",\"start\":1.82,\"end\":2.12}]"} {"utterance_id": "4507-16021-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0047.flac", "task_type": "semantic_detection", "query": "money", "prompt": "After listening to the whole recording, output every instance of a money or currency term with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.76, 10.37, "money"]], "response_mode": "offline", "answer": "[{\"word\":\"money\",\"start\":9.76,\"end\":10.37}]"} {"utterance_id": "1995-1837-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0014.flac", "task_type": "lexical_detection", "query": "dazzling", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"dazzling\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.86, 5.51, "dazzling"]], "response_mode": "offline", "answer": "[{\"word\":\"dazzling\",\"start\":4.86,\"end\":5.51}]"} {"utterance_id": "8463-287645-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0005.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.52, 3.16, "coachman"]], "response_mode": "offline", "answer": "[{\"word\":\"coachman\",\"start\":2.52,\"end\":3.16}]"} {"utterance_id": "5639-40744-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0003.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Once the audio has finished, list each occurrence of a person's name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[7.89, 8.47, "one"], [21.63, 22.41, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.89,\"end\":8.47},{\"word\":\"two\",\"start\":21.63,\"end\":22.41}]"} {"utterance_id": "2094-142345-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0006.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[12.84, 13.47, "nose"]], "response_mode": "offline", "answer": "[{\"word\":\"nose\",\"start\":12.84,\"end\":13.47}]"} {"utterance_id": "7729-102255-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0024.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Do not answer until the audio ends. Then list each instance of a state or province name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.51, 0.79, "state"]], "response_mode": "offline", "answer": "[{\"word\":\"state\",\"start\":0.51,\"end\":0.79}]"} {"utterance_id": "908-31957-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0022.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Once the audio has finished, identify every instance of a color name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[10.1, 10.5, "green"]], "response_mode": "offline", "answer": "[{\"word\":\"green\",\"start\":10.1,\"end\":10.5}]"} {"utterance_id": "8555-292519-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0007.flac", "task_type": "lexical_counting", "query": "heart", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"heart\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.88, 1.44, "one"], [6.65, 7.1, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.88,\"end\":1.44},{\"word\":\"two\",\"start\":6.65,\"end\":7.1}]"} {"utterance_id": "5683-32879-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0023.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.1, 1.64, "rachel"]], "response_mode": "offline", "answer": "[{\"word\":\"rachel\",\"start\":1.1,\"end\":1.64}]"} {"utterance_id": "4507-16021-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0041.flac", "task_type": "lexical_detection", "query": "dark", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"dark\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.28, 2.81, "dark"]], "response_mode": "offline", "answer": "[{\"word\":\"dark\",\"start\":2.28,\"end\":2.81}]"} {"utterance_id": "672-122797-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0045.flac", "task_type": "lexical_detection", "query": "trunks", "prompt": "After the audio ends, extract all occurrences of the word \"trunks\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.38, 11.95, "trunks"]], "response_mode": "offline", "answer": "[{\"word\":\"trunks\",\"start\":11.38,\"end\":11.95}]"} {"utterance_id": "4446-2273-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0028.flac", "task_type": "lexical_detection", "query": "sing", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"sing\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.08, 2.54, "sing"]], "response_mode": "offline", "answer": "[{\"word\":\"sing\",\"start\":2.08,\"end\":2.54}]"} {"utterance_id": "8455-210777-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0054.flac", "task_type": "lexical_detection", "query": "follows", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"follows\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.57, 2.23, "follows"]], "response_mode": "offline", "answer": "[{\"word\":\"follows\",\"start\":1.57,\"end\":2.23}]"} {"utterance_id": "8555-284449-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0017.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Process the full audio offline. Find every instance of a color name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.24, 10.4, "blue"]], "response_mode": "offline", "answer": "[{\"word\":\"blue\",\"start\":10.24,\"end\":10.4}]"} {"utterance_id": "6829-68769-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0047.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.45, 2.22, "mister marshall"]], "response_mode": "offline", "answer": "[{\"word\":\"mister marshall\",\"start\":1.45,\"end\":2.22}]"} {"utterance_id": "237-134500-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0012.flac", "task_type": "lexical_detection", "query": "cherries", "prompt": "Process the full audio offline. Find every instance of the word \"cherries\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.73, 2.24, "cherries"]], "response_mode": "offline", "answer": "[{\"word\":\"cherries\",\"start\":1.73,\"end\":2.24}]"} {"utterance_id": "7729-102255-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0016.flac", "task_type": "semantic_counting", "query": "state_or_province", "prompt": "After listening to the whole recording, output timed count words for a state or province name (one, two, three, ...) as JSON. If none, reply [].", "labels": [[4.81, 5.24, "one"], [6.32, 6.84, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.81,\"end\":5.24},{\"word\":\"two\",\"start\":6.32,\"end\":6.84}]"} {"utterance_id": "3729-6852-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0004.flac", "task_type": "lexical_detection", "query": "sympathy", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"sympathy\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.94, 5.61, "sympathy"]], "response_mode": "offline", "answer": "[{\"word\":\"sympathy\",\"start\":4.94,\"end\":5.61}]"} {"utterance_id": "672-122797-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0043.flac", "task_type": "lexical_detection", "query": "meaning", "prompt": "After listening to the whole recording, output every instance of the word \"meaning\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.09, 1.59, "meaning"]], "response_mode": "offline", "answer": "[{\"word\":\"meaning\",\"start\":1.09,\"end\":1.59}]"} {"utterance_id": "1221-135766-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0009.flac", "task_type": "lexical_detection", "query": "accordance", "prompt": "After listening to the whole recording, output every instance of the word \"accordance\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.54, 8.1, "accordance"]], "response_mode": "offline", "answer": "[{\"word\":\"accordance\",\"start\":7.54,\"end\":8.1}]"} {"utterance_id": "1188-133604-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0008.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "When the utterance is complete, report all instances of a plant name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.52, 8.89, "grass"]], "response_mode": "offline", "answer": "[{\"word\":\"grass\",\"start\":8.52,\"end\":8.89}]"} {"utterance_id": "6829-68769-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0046.flac", "task_type": "lexical_detection", "query": "foolish", "prompt": "Once the audio has finished, identify every instance of the word \"foolish\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.51, 0.94, "foolish"]], "response_mode": "offline", "answer": "[{\"word\":\"foolish\",\"start\":0.51,\"end\":0.94}]"} {"utterance_id": "4970-29095-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0018.flac", "task_type": "lexical_detection", "query": "sounding", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"sounding\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.86, 4.32, "sounding"]], "response_mode": "offline", "answer": "[{\"word\":\"sounding\",\"start\":3.86,\"end\":4.32}]"} {"utterance_id": "3729-6852-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0002.flac", "task_type": "lexical_detection", "query": "quote", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"quote\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.25, 6.6, "quote"]], "response_mode": "offline", "answer": "[{\"word\":\"quote\",\"start\":6.25,\"end\":6.6}]"} {"utterance_id": "8230-279154-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0009.flac", "task_type": "lexical_detection", "query": "bodily", "prompt": "After the audio ends, extract all occurrences of the word \"bodily\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.63, 14.36, "bodily"]], "response_mode": "offline", "answer": "[{\"word\":\"bodily\",\"start\":13.63,\"end\":14.36}]"} {"utterance_id": "2300-131720-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0032.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After the audio ends, extract all occurrences of an organization name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.87, 7.4, "company"]], "response_mode": "offline", "answer": "[{\"word\":\"company\",\"start\":6.87,\"end\":7.4}]"} {"utterance_id": "4970-29095-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0007.flac", "task_type": "lexical_detection", "query": "theirs", "prompt": "Process the full audio offline. Find every instance of the word \"theirs\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.85, 12.47, "theirs"]], "response_mode": "offline", "answer": "[{\"word\":\"theirs\",\"start\":11.85,\"end\":12.47}]"} {"utterance_id": "260-123286-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0028.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.6, 3.31, "naturalists"]], "response_mode": "offline", "answer": "[{\"word\":\"naturalists\",\"start\":2.6,\"end\":3.31}]"} {"utterance_id": "2830-3980-0074", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0074.flac", "task_type": "lexical_detection", "query": "developed", "prompt": "When the utterance is complete, report all instances of the word \"developed\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.56, 3.02, "developed"]], "response_mode": "offline", "answer": "[{\"word\":\"developed\",\"start\":2.56,\"end\":3.02}]"} {"utterance_id": "5142-36377-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0025.flac", "task_type": "lexical_detection", "query": "aside", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"aside\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[9.23, 9.73, "aside"]], "response_mode": "offline", "answer": "[{\"word\":\"aside\",\"start\":9.23,\"end\":9.73}]"} {"utterance_id": "7176-92135-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0029.flac", "task_type": "lexical_detection", "query": "want", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"want\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.32, 0.52, "want"]], "response_mode": "offline", "answer": "[{\"word\":\"want\",\"start\":0.32,\"end\":0.52}]"} {"utterance_id": "2094-142345-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0047.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.51, 0.93, "poyser"]], "response_mode": "offline", "answer": "[{\"word\":\"poyser\",\"start\":0.51,\"end\":0.93}]"} {"utterance_id": "1089-134686-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0020.flac", "task_type": "lexical_counting", "query": "pound", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"pound\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[1.6, 2.1, "one"], [3.99, 4.48, "two"], [9.2, 9.63, "three"], [11.28, 11.73, "four"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.6,\"end\":2.1},{\"word\":\"two\",\"start\":3.99,\"end\":4.48},{\"word\":\"three\",\"start\":9.2,\"end\":9.63},{\"word\":\"four\",\"start\":11.28,\"end\":11.73}]"} {"utterance_id": "2830-3980-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0041.flac", "task_type": "lexical_detection", "query": "remission", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"remission\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.45, 1.87, "remission"]], "response_mode": "offline", "answer": "[{\"word\":\"remission\",\"start\":1.45,\"end\":1.87}]"} {"utterance_id": "61-70968-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0023.flac", "task_type": "lexical_detection", "query": "played", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"played\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.16, 3.49, "played"]], "response_mode": "offline", "answer": "[{\"word\":\"played\",\"start\":3.16,\"end\":3.49}]"} {"utterance_id": "3729-6852-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0004.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.75, 1.34, "face"], [11.22, 11.84, "feature"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":0.75,\"end\":1.34},{\"word\":\"feature\",\"start\":11.22,\"end\":11.84}]"} {"utterance_id": "7729-102255-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0002.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Process the full audio offline. Find every instance of an organization name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.25, 7.88, "parties"]], "response_mode": "offline", "answer": "[{\"word\":\"parties\",\"start\":7.25,\"end\":7.88}]"} {"utterance_id": "908-157963-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0020.flac", "task_type": "lexical_detection", "query": "tender", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"tender\" with fields word, start, end (seconds). If none, reply [].", "labels": [[8.31, 8.69, "tender"]], "response_mode": "offline", "answer": "[{\"word\":\"tender\",\"start\":8.31,\"end\":8.69}]"} {"utterance_id": "3570-5694-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0002.flac", "task_type": "lexical_counting", "query": "consumption", "prompt": "After the audio ends, tally every instance of the word \"consumption\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.56, 1.13, "one"], [6.94, 7.54, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.56,\"end\":1.13},{\"word\":\"two\",\"start\":6.94,\"end\":7.54}]"} {"utterance_id": "260-123288-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0014.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.45, 0.87, "hans"]], "response_mode": "offline", "answer": "[{\"word\":\"hans\",\"start\":0.45,\"end\":0.87}]"} {"utterance_id": "4992-41797-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0002.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.36, 3.79, "doctor"]], "response_mode": "offline", "answer": "[{\"word\":\"doctor\",\"start\":3.36,\"end\":3.79}]"} {"utterance_id": "7729-102255-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0037.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.29, 0.57, "half"]], "response_mode": "offline", "answer": "[{\"word\":\"half\",\"start\":0.29,\"end\":0.57}]"} {"utterance_id": "8455-210777-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0044.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Process the full audio offline. Find every instance of a country name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.29, 2.71, "england"]], "response_mode": "offline", "answer": "[{\"word\":\"england\",\"start\":2.29,\"end\":2.71}]"} {"utterance_id": "7127-75947-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0021.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.38, 0.98, "la valliere"]], "response_mode": "offline", "answer": "[{\"word\":\"la valliere\",\"start\":0.38,\"end\":0.98}]"} {"utterance_id": "8224-274384-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0012.flac", "task_type": "lexical_detection", "query": "tears", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"tears\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.66, 14.29, "tears"]], "response_mode": "offline", "answer": "[{\"word\":\"tears\",\"start\":13.66,\"end\":14.29}]"} {"utterance_id": "7021-85628-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0011.flac", "task_type": "semantic_detection", "query": "color", "prompt": "When the utterance is complete, report all instances of a color name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.03, 2.28, "blue"], [6.88, 7.4, "golden"]], "response_mode": "offline", "answer": "[{\"word\":\"blue\",\"start\":2.03,\"end\":2.28},{\"word\":\"golden\",\"start\":6.88,\"end\":7.4}]"} {"utterance_id": "5142-33396-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0004.flac", "task_type": "lexical_detection", "query": "brothers", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"brothers\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.4, 1.99, "brothers"]], "response_mode": "offline", "answer": "[{\"word\":\"brothers\",\"start\":1.4,\"end\":1.99}]"} {"utterance_id": "6829-68769-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0028.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an organization name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.55, 3.1, "concern"]], "response_mode": "offline", "answer": "[{\"word\":\"concern\",\"start\":2.55,\"end\":3.1}]"} {"utterance_id": "5105-28241-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0013.flac", "task_type": "lexical_detection", "query": "visible", "prompt": "After the audio ends, extract all occurrences of the word \"visible\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.97, 4.47, "visible"]], "response_mode": "offline", "answer": "[{\"word\":\"visible\",\"start\":3.97,\"end\":4.47}]"} {"utterance_id": "6829-68771-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0016.flac", "task_type": "lexical_detection", "query": "favorite", "prompt": "After listening to the whole recording, output every instance of the word \"favorite\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.37, 6.88, "favorite"]], "response_mode": "offline", "answer": "[{\"word\":\"favorite\",\"start\":6.37,\"end\":6.88}]"} {"utterance_id": "8555-284447-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0011.flac", "task_type": "lexical_detection", "query": "suppose", "prompt": "After listening to the whole recording, output every instance of the word \"suppose\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 0.98, "suppose"]], "response_mode": "offline", "answer": "[{\"word\":\"suppose\",\"start\":0.42,\"end\":0.98}]"} {"utterance_id": "3570-5694-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0005.flac", "task_type": "lexical_counting", "query": "certain", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"certain\" (word, start, end). If none, reply [].", "labels": [[1.22, 1.65, "one"], [3.6, 4.06, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.22,\"end\":1.65},{\"word\":\"two\",\"start\":3.6,\"end\":4.06}]"} {"utterance_id": "260-123440-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0002.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Process the full audio offline. Find every instance of a color name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.8, 1.12, "white"], [3.75, 4.1, "white"]], "response_mode": "offline", "answer": "[{\"word\":\"white\",\"start\":0.8,\"end\":1.12},{\"word\":\"white\",\"start\":3.75,\"end\":4.1}]"} {"utterance_id": "2961-961-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0006.flac", "task_type": "lexical_detection", "query": "poem", "prompt": "Once the audio has finished, identify every instance of the word \"poem\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.87, 2.3, "poem"]], "response_mode": "offline", "answer": "[{\"word\":\"poem\",\"start\":1.87,\"end\":2.3}]"} {"utterance_id": "6930-76324-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0006.flac", "task_type": "lexical_detection", "query": "hers", "prompt": "After the audio ends, extract all occurrences of the word \"hers\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.18, 0.45, "hers"]], "response_mode": "offline", "answer": "[{\"word\":\"hers\",\"start\":0.18,\"end\":0.45}]"} {"utterance_id": "908-157963-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0029.flac", "task_type": "lexical_detection", "query": "wind", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"wind\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.91, 3.4, "wind"]], "response_mode": "offline", "answer": "[{\"word\":\"wind\",\"start\":2.91,\"end\":3.4}]"} {"utterance_id": "1221-135767-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0019.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.21, 0.58, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":0.21,\"end\":0.58}]"} {"utterance_id": "5639-40744-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0034.flac", "task_type": "lexical_detection", "query": "admiration", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"admiration\" with fields word, start, end (seconds). If none, reply [].", "labels": [[8.65, 9.41, "admiration"]], "response_mode": "offline", "answer": "[{\"word\":\"admiration\",\"start\":8.65,\"end\":9.41}]"} {"utterance_id": "8463-294828-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0004.flac", "task_type": "lexical_detection", "query": "hold", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"hold\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.24, 1.53, "hold"]], "response_mode": "offline", "answer": "[{\"word\":\"hold\",\"start\":1.24,\"end\":1.53}]"} {"utterance_id": "8230-279154-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0018.flac", "task_type": "lexical_counting", "query": "knowing", "prompt": "Once the audio has finished, list each occurrence of the word \"knowing\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[3.23, 3.7, "one"], [8.08, 8.61, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.23,\"end\":3.7},{\"word\":\"two\",\"start\":8.08,\"end\":8.61}]"} {"utterance_id": "4507-16021-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0020.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.13, 13.25, "trades professions"]], "response_mode": "offline", "answer": "[{\"word\":\"trades professions\",\"start\":11.13,\"end\":13.25}]"} {"utterance_id": "1188-133604-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0026.flac", "task_type": "lexical_detection", "query": "sunshine", "prompt": "Process the full audio offline. Find every instance of the word \"sunshine\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[16.15, 16.76, "sunshine"]], "response_mode": "offline", "answer": "[{\"word\":\"sunshine\",\"start\":16.15,\"end\":16.76}]"} {"utterance_id": "237-134493-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0015.flac", "task_type": "lexical_detection", "query": "unusual", "prompt": "Once the audio has finished, identify every instance of the word \"unusual\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.55, 4.27, "unusual"]], "response_mode": "offline", "answer": "[{\"word\":\"unusual\",\"start\":3.55,\"end\":4.27}]"} {"utterance_id": "5683-32866-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0024.flac", "task_type": "lexical_detection", "query": "trouble", "prompt": "Process the full audio offline. Find every instance of the word \"trouble\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.03, 1.34, "trouble"]], "response_mode": "offline", "answer": "[{\"word\":\"trouble\",\"start\":1.03,\"end\":1.34}]"} {"utterance_id": "8455-210777-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.8, 2.23, "eva"]], "response_mode": "offline", "answer": "[{\"word\":\"eva\",\"start\":1.8,\"end\":2.23}]"} {"utterance_id": "4077-13754-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0008.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a number (word, start, end). If none, reply [].", "labels": [[3.24, 3.46, "one"], [7.57, 7.94, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.24,\"end\":3.46},{\"word\":\"two\",\"start\":7.57,\"end\":7.94}]"} {"utterance_id": "1580-141084-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0006.flac", "task_type": "lexical_detection", "query": "position", "prompt": "Once the audio has finished, identify every instance of the word \"position\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.27, 1.92, "position"]], "response_mode": "offline", "answer": "[{\"word\":\"position\",\"start\":1.27,\"end\":1.92}]"} {"utterance_id": "1320-122617-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0000.flac", "task_type": "lexical_detection", "query": "difficulties", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"difficulties\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.85, 5.58, "difficulties"]], "response_mode": "offline", "answer": "[{\"word\":\"difficulties\",\"start\":4.85,\"end\":5.58}]"} {"utterance_id": "4507-16021-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0008.flac", "task_type": "lexical_detection", "query": "does", "prompt": "After the audio ends, extract all occurrences of the word \"does\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.72, 3.3, "does"]], "response_mode": "offline", "answer": "[{\"word\":\"does\",\"start\":2.72,\"end\":3.3}]"} {"utterance_id": "7176-92135-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0020.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.63, 6.6, "horatio hamlet"]], "response_mode": "offline", "answer": "[{\"word\":\"horatio hamlet\",\"start\":5.63,\"end\":6.6}]"} {"utterance_id": "2830-3980-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0054.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "You will hear a complete utterance. When it ends, list every instance of a family relation you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.84, 6.32, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":5.84,\"end\":6.32}]"} {"utterance_id": "7021-85628-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0015.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen carefully to the entire clip, then report every instance of a food name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.63, 5.2, "cakes"]], "response_mode": "offline", "answer": "[{\"word\":\"cakes\",\"start\":4.63,\"end\":5.2}]"} {"utterance_id": "1995-1826-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0019.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a plant name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.24, 0.77, "cotton"]], "response_mode": "offline", "answer": "[{\"word\":\"cotton\",\"start\":0.24,\"end\":0.77}]"} {"utterance_id": "8463-287645-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0007.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "After the audio ends, tally every instance of a number. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[3.41, 3.78, "one"], [5.93, 6.14, "two"], [19.35, 19.61, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.41,\"end\":3.78},{\"word\":\"two\",\"start\":5.93,\"end\":6.14},{\"word\":\"three\",\"start\":19.35,\"end\":19.61}]"} {"utterance_id": "5142-33396-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0047.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.75, 2.35, "fists"]], "response_mode": "offline", "answer": "[{\"word\":\"fists\",\"start\":1.75,\"end\":2.35}]"} {"utterance_id": "2830-3980-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.89, 3.52, "saint paul"]], "response_mode": "offline", "answer": "[{\"word\":\"saint paul\",\"start\":2.89,\"end\":3.52}]"} {"utterance_id": "5142-33396-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0006.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.87, 1.28, "twenty"], [4.9, 5.35, "fifteen"]], "response_mode": "offline", "answer": "[{\"word\":\"twenty\",\"start\":0.87,\"end\":1.28},{\"word\":\"fifteen\",\"start\":4.9,\"end\":5.35}]"} {"utterance_id": "3570-5696-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0008.flac", "task_type": "lexical_detection", "query": "grounds", "prompt": "When the utterance is complete, report all instances of the word \"grounds\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.91, 1.32, "grounds"]], "response_mode": "offline", "answer": "[{\"word\":\"grounds\",\"start\":0.91,\"end\":1.32}]"} {"utterance_id": "5683-32866-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0011.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.3, 6.57, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":6.3,\"end\":6.57}]"} {"utterance_id": "1284-134647-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0005.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After the audio ends, extract all occurrences of an organization name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.62, 14.43, "catholic church"]], "response_mode": "offline", "answer": "[{\"word\":\"catholic church\",\"start\":13.62,\"end\":14.43}]"} {"utterance_id": "6930-81414-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0003.flac", "task_type": "lexical_detection", "query": "sound", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"sound\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.75, 1.31, "sound"]], "response_mode": "offline", "answer": "[{\"word\":\"sound\",\"start\":0.75,\"end\":1.31}]"} {"utterance_id": "6829-68769-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0002.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.56, 2.96, "lawyer"]], "response_mode": "offline", "answer": "[{\"word\":\"lawyer\",\"start\":2.56,\"end\":2.96}]"} {"utterance_id": "4446-2273-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0003.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.71, 1.05, "bartley"], [3.33, 3.84, "marie"]], "response_mode": "offline", "answer": "[{\"word\":\"bartley\",\"start\":0.71,\"end\":1.05},{\"word\":\"marie\",\"start\":3.33,\"end\":3.84}]"} {"utterance_id": "1284-1181-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0011.flac", "task_type": "semantic_counting", "query": "color", "prompt": "Process the full audio offline. For every occurrence of a color name, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.6, 5.06, "one"], [6.59, 6.97, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.6,\"end\":5.06},{\"word\":\"two\",\"start\":6.59,\"end\":6.97}]"} {"utterance_id": "7176-88083-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0011.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.7, 3.07, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":2.7,\"end\":3.07}]"} {"utterance_id": "8463-294828-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0031.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.74, 7.05, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":6.74,\"end\":7.05}]"} {"utterance_id": "237-134493-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0011.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.4, 3.85, "athlete"]], "response_mode": "offline", "answer": "[{\"word\":\"athlete\",\"start\":3.4,\"end\":3.85}]"} {"utterance_id": "1089-134691-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0020.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a person's name: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[1.01, 1.94, "one"], [3.25, 3.86, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.01,\"end\":1.94},{\"word\":\"two\",\"start\":3.25,\"end\":3.86}]"} {"utterance_id": "61-70968-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0003.flac", "task_type": "lexical_detection", "query": "yet", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"yet\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.69, 2.85, "yet"]], "response_mode": "offline", "answer": "[{\"word\":\"yet\",\"start\":2.69,\"end\":2.85}]"} {"utterance_id": "3575-170457-0056", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0056.flac", "task_type": "lexical_detection", "query": "whether", "prompt": "Once the audio has finished, identify every instance of the word \"whether\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.56, 0.85, "whether"]], "response_mode": "offline", "answer": "[{\"word\":\"whether\",\"start\":0.56,\"end\":0.85}]"} {"utterance_id": "5142-36600-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36600/5142-36600-0000.flac", "task_type": "lexical_detection", "query": "chapter", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"chapter\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.16, 0.6, "chapter"]], "response_mode": "offline", "answer": "[{\"word\":\"chapter\",\"start\":0.16,\"end\":0.6}]"} {"utterance_id": "7176-92135-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0028.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the full audio. After it finishes, find every instance of a plant name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.17, 0.72, "larkspur"]], "response_mode": "offline", "answer": "[{\"word\":\"larkspur\",\"start\":0.17,\"end\":0.72}]"} {"utterance_id": "3570-5695-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0001.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the full audio. After it finishes, find every instance of an organization name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.66, 8.35, "institution"]], "response_mode": "offline", "answer": "[{\"word\":\"institution\",\"start\":7.66,\"end\":8.35}]"} {"utterance_id": "1995-1826-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0014.flac", "task_type": "lexical_detection", "query": "paused", "prompt": "After listening to the whole recording, output every instance of the word \"paused\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.76, 2.32, "paused"]], "response_mode": "offline", "answer": "[{\"word\":\"paused\",\"start\":1.76,\"end\":2.32}]"} {"utterance_id": "4992-23283-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0012.flac", "task_type": "lexical_counting", "query": "man", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"man\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[4.21, 4.62, "one"], [17.73, 18.06, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.21,\"end\":4.62},{\"word\":\"two\",\"start\":17.73,\"end\":18.06}]"} {"utterance_id": "5142-36377-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0004.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the full audio. After it finishes, find every instance of a family relation. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.96, 5.43, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":4.96,\"end\":5.43}]"} {"utterance_id": "1995-1837-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0024.flac", "task_type": "lexical_detection", "query": "bird", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"bird\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.18, 4.5, "bird"]], "response_mode": "offline", "answer": "[{\"word\":\"bird\",\"start\":4.18,\"end\":4.5}]"} {"utterance_id": "7176-88083-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0011.flac", "task_type": "lexical_detection", "query": "baffled", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"baffled\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.47, 7.87, "baffled"]], "response_mode": "offline", "answer": "[{\"word\":\"baffled\",\"start\":7.47,\"end\":7.87}]"} {"utterance_id": "5639-40744-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0014.flac", "task_type": "lexical_detection", "query": "cabinet", "prompt": "Once the audio has finished, identify every instance of the word \"cabinet\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.26, 6.8, "cabinet"]], "response_mode": "offline", "answer": "[{\"word\":\"cabinet\",\"start\":6.26,\"end\":6.8}]"} {"utterance_id": "7176-88083-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0025.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.75, 5.07, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":4.75,\"end\":5.07}]"} {"utterance_id": "2830-3980-0068", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0068.flac", "task_type": "semantic_detection", "query": "money", "prompt": "After listening to the whole recording, output every instance of a money or currency term with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.3, 1.81, "ransom"]], "response_mode": "offline", "answer": "[{\"word\":\"ransom\",\"start\":1.3,\"end\":1.81}]"} {"utterance_id": "4507-16021-0058", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0058.flac", "task_type": "lexical_detection", "query": "enemy", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"enemy\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.62, 2.09, "enemy"]], "response_mode": "offline", "answer": "[{\"word\":\"enemy\",\"start\":1.62,\"end\":2.09}]"} {"utterance_id": "1284-134647-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0000.flac", "task_type": "lexical_detection", "query": "grateful", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"grateful\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.63, 1.09, "grateful"]], "response_mode": "offline", "answer": "[{\"word\":\"grateful\",\"start\":0.63,\"end\":1.09}]"} {"utterance_id": "5105-28240-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0016.flac", "task_type": "lexical_detection", "query": "responded", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"responded\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.31, 2.93, "responded"]], "response_mode": "offline", "answer": "[{\"word\":\"responded\",\"start\":2.31,\"end\":2.93}]"} {"utterance_id": "4446-2271-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0018.flac", "task_type": "lexical_detection", "query": "considered", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"considered\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.56, 1.04, "considered"]], "response_mode": "offline", "answer": "[{\"word\":\"considered\",\"start\":0.56,\"end\":1.04}]"} {"utterance_id": "672-122797-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0026.flac", "task_type": "lexical_detection", "query": "silken", "prompt": "When the utterance is complete, report all instances of the word \"silken\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.27, 4.01, "silken"]], "response_mode": "offline", "answer": "[{\"word\":\"silken\",\"start\":3.27,\"end\":4.01}]"} {"utterance_id": "1320-122617-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0016.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[5.79, 6.24, "david"]], "response_mode": "offline", "answer": "[{\"word\":\"david\",\"start\":5.79,\"end\":6.24}]"} {"utterance_id": "4446-2273-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0016.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "After listening to the whole recording, output timed count words for a body part (one, two, three, ...) as JSON. If none, reply [].", "labels": [[0.42, 0.65, "one"], [3.02, 3.37, "two"], [6.77, 7.18, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.42,\"end\":0.65},{\"word\":\"two\",\"start\":3.02,\"end\":3.37},{\"word\":\"three\",\"start\":6.77,\"end\":7.18}]"} {"utterance_id": "1221-135767-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0010.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.26, 5.73, "hearts"]], "response_mode": "offline", "answer": "[{\"word\":\"hearts\",\"start\":5.26,\"end\":5.73}]"} {"utterance_id": "5105-28240-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0023.flac", "task_type": "lexical_detection", "query": "diminution", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"diminution\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.84, 1.52, "diminution"]], "response_mode": "offline", "answer": "[{\"word\":\"diminution\",\"start\":0.84,\"end\":1.52}]"} {"utterance_id": "1089-134686-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0023.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.74, 8.1, "body"]], "response_mode": "offline", "answer": "[{\"word\":\"body\",\"start\":7.74,\"end\":8.1}]"} {"utterance_id": "1995-1837-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0017.flac", "task_type": "lexical_detection", "query": "perplexed", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"perplexed\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.17, 1.82, "perplexed"]], "response_mode": "offline", "answer": "[{\"word\":\"perplexed\",\"start\":1.17,\"end\":1.82}]"} {"utterance_id": "5683-32879-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0022.flac", "task_type": "lexical_detection", "query": "sure", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"sure\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.36, 2.72, "sure"]], "response_mode": "offline", "answer": "[{\"word\":\"sure\",\"start\":2.36,\"end\":2.72}]"} {"utterance_id": "2094-142345-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0022.flac", "task_type": "lexical_detection", "query": "indeed", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"indeed\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.36, 1.91, "indeed"]], "response_mode": "offline", "answer": "[{\"word\":\"indeed\",\"start\":1.36,\"end\":1.91}]"} {"utterance_id": "2094-142345-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0033.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.92, 10.32, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":9.92,\"end\":10.32}]"} {"utterance_id": "8555-292519-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0012.flac", "task_type": "lexical_detection", "query": "rain", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"rain\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.87, 2.41, "rain"]], "response_mode": "offline", "answer": "[{\"word\":\"rain\",\"start\":1.87,\"end\":2.41}]"} {"utterance_id": "7127-75947-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0024.flac", "task_type": "lexical_detection", "query": "see", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"see\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.64, 1.75, "see"]], "response_mode": "offline", "answer": "[{\"word\":\"see\",\"start\":1.64,\"end\":1.75}]"} {"utterance_id": "1995-1837-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0006.flac", "task_type": "lexical_detection", "query": "veiled", "prompt": "Process the full audio offline. Find every instance of the word \"veiled\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.25, 1.77, "veiled"]], "response_mode": "offline", "answer": "[{\"word\":\"veiled\",\"start\":1.25,\"end\":1.77}]"} {"utterance_id": "7176-92135-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0017.flac", "task_type": "lexical_detection", "query": "badly", "prompt": "When the utterance is complete, report all instances of the word \"badly\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.05, 1.56, "badly"]], "response_mode": "offline", "answer": "[{\"word\":\"badly\",\"start\":1.05,\"end\":1.56}]"} {"utterance_id": "7021-85628-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0024.flac", "task_type": "semantic_detection", "query": "color", "prompt": "You will hear a complete utterance. When it ends, list every instance of a color name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.46, 1.9, "red"]], "response_mode": "offline", "answer": "[{\"word\":\"red\",\"start\":1.46,\"end\":1.9}]"} {"utterance_id": "6930-81414-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0015.flac", "task_type": "lexical_detection", "query": "bewildered", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"bewildered\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.59, 3.51, "bewildered"]], "response_mode": "offline", "answer": "[{\"word\":\"bewildered\",\"start\":2.59,\"end\":3.51}]"} {"utterance_id": "121-127105-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0000.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "When the utterance is complete, report all instances of a time of day including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.03, 5.52, "evening"]], "response_mode": "offline", "answer": "[{\"word\":\"evening\",\"start\":5.03,\"end\":5.52}]"} {"utterance_id": "3575-170457-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0021.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.21, 3.82, "governess"]], "response_mode": "offline", "answer": "[{\"word\":\"governess\",\"start\":3.21,\"end\":3.82}]"} {"utterance_id": "1284-1180-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0007.flac", "task_type": "lexical_detection", "query": "crooked", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"crooked\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.03, 2.41, "crooked"]], "response_mode": "offline", "answer": "[{\"word\":\"crooked\",\"start\":2.03,\"end\":2.41}]"} {"utterance_id": "8230-279154-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0032.flac", "task_type": "lexical_detection", "query": "theory", "prompt": "Once the audio has finished, identify every instance of the word \"theory\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.41, 2.83, "theory"]], "response_mode": "offline", "answer": "[{\"word\":\"theory\",\"start\":2.41,\"end\":2.83}]"} {"utterance_id": "8230-279154-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0000.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Do not answer until the audio ends. Then list each instance of an ordinal number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.49, 4.12, "thirteenth"]], "response_mode": "offline", "answer": "[{\"word\":\"thirteenth\",\"start\":3.49,\"end\":4.12}]"} {"utterance_id": "2094-142345-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0042.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.41, 0.63, "hanna"]], "response_mode": "offline", "answer": "[{\"word\":\"hanna\",\"start\":0.41,\"end\":0.63}]"} {"utterance_id": "1089-134691-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0008.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "When the utterance is complete, report how a body part accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.01, 1.49, "one"], [14.05, 14.55, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.01,\"end\":1.49},{\"word\":\"two\",\"start\":14.05,\"end\":14.55}]"} {"utterance_id": "7127-75946-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0015.flac", "task_type": "lexical_detection", "query": "accompanied", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"accompanied\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.47, 5.04, "accompanied"]], "response_mode": "offline", "answer": "[{\"word\":\"accompanied\",\"start\":4.47,\"end\":5.04}]"} {"utterance_id": "5105-28240-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0011.flac", "task_type": "lexical_detection", "query": "finding", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"finding\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.93, 4.39, "finding"]], "response_mode": "offline", "answer": "[{\"word\":\"finding\",\"start\":3.93,\"end\":4.39}]"} {"utterance_id": "5639-40744-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0018.flac", "task_type": "lexical_detection", "query": "liable", "prompt": "Process the full audio offline. Find every instance of the word \"liable\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.26, 4.73, "liable"]], "response_mode": "offline", "answer": "[{\"word\":\"liable\",\"start\":4.26,\"end\":4.73}]"} {"utterance_id": "2094-142345-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0054.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.23, 2.9, "missus poyser"]], "response_mode": "offline", "answer": "[{\"word\":\"missus poyser\",\"start\":2.23,\"end\":2.9}]"} {"utterance_id": "6930-75918-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0012.flac", "task_type": "lexical_detection", "query": "annoyance", "prompt": "After listening to the whole recording, output every instance of the word \"annoyance\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.77, 1.37, "annoyance"]], "response_mode": "offline", "answer": "[{\"word\":\"annoyance\",\"start\":0.77,\"end\":1.37}]"} {"utterance_id": "4507-16021-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0050.flac", "task_type": "lexical_detection", "query": "belong", "prompt": "After listening to the whole recording, output every instance of the word \"belong\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.79, 1.18, "belong"]], "response_mode": "offline", "answer": "[{\"word\":\"belong\",\"start\":0.79,\"end\":1.18}]"} {"utterance_id": "1188-133604-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0023.flac", "task_type": "lexical_counting", "query": "black", "prompt": "After the audio ends, tally every instance of the word \"black\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[8.61, 9.16, "one"], [14.77, 15.24, "two"], [20.55, 20.84, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":8.61,\"end\":9.16},{\"word\":\"two\",\"start\":14.77,\"end\":15.24},{\"word\":\"three\",\"start\":20.55,\"end\":20.84}]"} {"utterance_id": "4077-13754-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0000.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, list every instance of an organization name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 0.84, "army"]], "response_mode": "offline", "answer": "[{\"word\":\"army\",\"start\":0.51,\"end\":0.84}]"} {"utterance_id": "1320-122617-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0003.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.04, 3.55, "scout"]], "response_mode": "offline", "answer": "[{\"word\":\"scout\",\"start\":3.04,\"end\":3.55}]"} {"utterance_id": "4077-13754-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0008.flac", "task_type": "semantic_counting", "query": "country", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a country name: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[3.46, 3.99, "one"], [7.0, 7.4, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.46,\"end\":3.99},{\"word\":\"two\",\"start\":7.0,\"end\":7.4}]"} {"utterance_id": "3570-5694-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0000.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen carefully to the entire clip, then report every instance of a money or currency term as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.84, 8.4, "pecuniary"]], "response_mode": "offline", "answer": "[{\"word\":\"pecuniary\",\"start\":7.84,\"end\":8.4}]"} {"utterance_id": "4446-2275-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0031.flac", "task_type": "lexical_detection", "query": "understand", "prompt": "Process the full audio offline. Find every instance of the word \"understand\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.73, 1.24, "understand"]], "response_mode": "offline", "answer": "[{\"word\":\"understand\",\"start\":0.73,\"end\":1.24}]"} {"utterance_id": "1188-133604-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0037.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.56, 1.98, "master"]], "response_mode": "offline", "answer": "[{\"word\":\"master\",\"start\":1.56,\"end\":1.98}]"} {"utterance_id": "1995-1837-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0023.flac", "task_type": "lexical_detection", "query": "creeping", "prompt": "After the audio ends, extract all occurrences of the word \"creeping\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.34, 3.76, "creeping"], [11.5, 11.93, "creeping"]], "response_mode": "offline", "answer": "[{\"word\":\"creeping\",\"start\":3.34,\"end\":3.76},{\"word\":\"creeping\",\"start\":11.5,\"end\":11.93}]"} {"utterance_id": "4077-13751-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0011.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Listen carefully to the entire clip, then report every instance of a state or province name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.39, 3.93, "missouri"], [5.41, 6.22, "jackson county"]], "response_mode": "offline", "answer": "[{\"word\":\"missouri\",\"start\":3.39,\"end\":3.93},{\"word\":\"jackson county\",\"start\":5.41,\"end\":6.22}]"} {"utterance_id": "5142-36377-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0008.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Listen to the full audio. Count every occurrence of a number. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.38, 2.6, "one"], [14.47, 14.68, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.38,\"end\":2.6},{\"word\":\"two\",\"start\":14.47,\"end\":14.68}]"} {"utterance_id": "7021-85628-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0016.flac", "task_type": "lexical_detection", "query": "fine", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"fine\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.77, 2.27, "fine"]], "response_mode": "offline", "answer": "[{\"word\":\"fine\",\"start\":1.77,\"end\":2.27}]"} {"utterance_id": "2300-131720-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.28, 0.79, "edison"]], "response_mode": "offline", "answer": "[{\"word\":\"edison\",\"start\":0.28,\"end\":0.79}]"} {"utterance_id": "5142-33396-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0000.flac", "task_type": "lexical_detection", "query": "asked", "prompt": "When the utterance is complete, report all instances of the word \"asked\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.47, 1.87, "asked"]], "response_mode": "offline", "answer": "[{\"word\":\"asked\",\"start\":1.47,\"end\":1.87}]"} {"utterance_id": "8463-294828-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0028.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.11, 5.74, "heels"]], "response_mode": "offline", "answer": "[{\"word\":\"heels\",\"start\":5.11,\"end\":5.74}]"} {"utterance_id": "672-122797-0058", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0058.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "When the utterance is complete, report all instances of an animal name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.61, 4.16, "mice"]], "response_mode": "offline", "answer": "[{\"word\":\"mice\",\"start\":3.61,\"end\":4.16}]"} {"utterance_id": "4992-23283-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0000.flac", "task_type": "lexical_detection", "query": "powerful", "prompt": "Process the full audio offline. Find every instance of the word \"powerful\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.34, 3.95, "powerful"]], "response_mode": "offline", "answer": "[{\"word\":\"powerful\",\"start\":3.34,\"end\":3.95}]"} {"utterance_id": "1188-133604-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0003.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Process the full audio offline. Find every instance of a color name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.54, 10.98, "hue"]], "response_mode": "offline", "answer": "[{\"word\":\"hue\",\"start\":10.54,\"end\":10.98}]"} {"utterance_id": "1320-122617-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0001.flac", "task_type": "lexical_detection", "query": "intently", "prompt": "After the audio ends, extract all occurrences of the word \"intently\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.44, 5.02, "intently"]], "response_mode": "offline", "answer": "[{\"word\":\"intently\",\"start\":4.44,\"end\":5.02}]"} {"utterance_id": "7176-92135-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0013.flac", "task_type": "lexical_detection", "query": "see", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"see\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.59, 1.81, "see"]], "response_mode": "offline", "answer": "[{\"word\":\"see\",\"start\":1.59,\"end\":1.81}]"} {"utterance_id": "2094-142345-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0059.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen to the full audio. After it finishes, find every instance of a food name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.29, 5.51, "milk"]], "response_mode": "offline", "answer": "[{\"word\":\"milk\",\"start\":5.29,\"end\":5.51}]"} {"utterance_id": "5142-36586-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36586/5142-36586-0002.flac", "task_type": "lexical_detection", "query": "variability", "prompt": "Once the audio has finished, identify every instance of the word \"variability\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.37, 1.03, "variability"]], "response_mode": "offline", "answer": "[{\"word\":\"variability\",\"start\":0.37,\"end\":1.03}]"} {"utterance_id": "6829-68771-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0026.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.53, 1.84, "head"]], "response_mode": "offline", "answer": "[{\"word\":\"head\",\"start\":1.53,\"end\":1.84}]"} {"utterance_id": "6829-68771-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.1, 4.83, "mister hopkins"], [6.11, 6.61, "forbes's"]], "response_mode": "offline", "answer": "[{\"word\":\"mister hopkins\",\"start\":4.1,\"end\":4.83},{\"word\":\"forbes's\",\"start\":6.11,\"end\":6.61}]"} {"utterance_id": "237-134500-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0011.flac", "task_type": "lexical_detection", "query": "grass", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"grass\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.71, 3.22, "grass"]], "response_mode": "offline", "answer": "[{\"word\":\"grass\",\"start\":2.71,\"end\":3.22}]"} {"utterance_id": "2830-3980-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0011.flac", "task_type": "lexical_detection", "query": "false", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"false\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.26, 1.62, "false"]], "response_mode": "offline", "answer": "[{\"word\":\"false\",\"start\":1.26,\"end\":1.62}]"} {"utterance_id": "237-126133-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0015.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "When the utterance is complete, report how a person's name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.93, 3.48, "one"], [8.33, 8.81, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.93,\"end\":3.48},{\"word\":\"two\",\"start\":8.33,\"end\":8.81}]"} {"utterance_id": "5105-28233-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28233/5105-28233-0010.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[11.16, 11.61, "trooper's"]], "response_mode": "offline", "answer": "[{\"word\":\"trooper's\",\"start\":11.16,\"end\":11.61}]"} {"utterance_id": "4077-13754-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0002.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[7.62, 8.23, "officials"]], "response_mode": "offline", "answer": "[{\"word\":\"officials\",\"start\":7.62,\"end\":8.23}]"} {"utterance_id": "2961-961-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0014.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a number with word/start/end. If none, reply [].", "labels": [[0.46, 1.55, "one"], [5.11, 6.03, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.46,\"end\":1.55},{\"word\":\"two\",\"start\":5.11,\"end\":6.03}]"} {"utterance_id": "672-122797-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0021.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.8, 1.96, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.8,\"end\":1.96}]"} {"utterance_id": "8455-210777-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0004.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.99, 2.57, "president"]], "response_mode": "offline", "answer": "[{\"word\":\"president\",\"start\":1.99,\"end\":2.57}]"} {"utterance_id": "8463-294825-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0010.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen carefully to the entire clip, then report every instance of an ordinal number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.92, 1.29, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":0.92,\"end\":1.29}]"} {"utterance_id": "7729-102255-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0010.flac", "task_type": "lexical_detection", "query": "extended", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"extended\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.84, 5.3, "extended"]], "response_mode": "offline", "answer": "[{\"word\":\"extended\",\"start\":4.84,\"end\":5.3}]"} {"utterance_id": "1089-134686-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0034.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.65, 4.05, "hands"]], "response_mode": "offline", "answer": "[{\"word\":\"hands\",\"start\":3.65,\"end\":4.05}]"} {"utterance_id": "908-31957-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0015.flac", "task_type": "lexical_counting", "query": "love", "prompt": "Once the audio has finished, list each occurrence of the word \"love\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[1.59, 2.05, "one"], [18.72, 19.28, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.59,\"end\":2.05},{\"word\":\"two\",\"start\":18.72,\"end\":19.28}]"} {"utterance_id": "1221-135767-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0004.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "When the utterance is complete, report how a body part accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.84, 9.21, "one"], [12.71, 13.1, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":8.84,\"end\":9.21},{\"word\":\"two\",\"start\":12.71,\"end\":13.1}]"} {"utterance_id": "8463-294828-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0030.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.8, 2.4, "commander farragut"]], "response_mode": "offline", "answer": "[{\"word\":\"commander farragut\",\"start\":0.8,\"end\":2.4}]"} {"utterance_id": "1284-134647-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0007.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, list every instance of an organization name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.09, 4.85, "donatists"]], "response_mode": "offline", "answer": "[{\"word\":\"donatists\",\"start\":4.09,\"end\":4.85}]"} {"utterance_id": "6829-68769-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0021.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen carefully to the entire clip, then report every instance of a color name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.47, 3.77, "gray"]], "response_mode": "offline", "answer": "[{\"word\":\"gray\",\"start\":3.47,\"end\":3.77}]"} {"utterance_id": "3570-5695-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0013.flac", "task_type": "lexical_detection", "query": "larger", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"larger\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.35, 1.69, "larger"]], "response_mode": "offline", "answer": "[{\"word\":\"larger\",\"start\":1.35,\"end\":1.69}]"} {"utterance_id": "1320-122617-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0020.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.52, 1.83, "david"]], "response_mode": "offline", "answer": "[{\"word\":\"david\",\"start\":1.52,\"end\":1.83}]"} {"utterance_id": "1284-134647-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0007.flac", "task_type": "lexical_detection", "query": "donatists", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"donatists\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.09, 4.85, "donatists"]], "response_mode": "offline", "answer": "[{\"word\":\"donatists\",\"start\":4.09,\"end\":4.85}]"} {"utterance_id": "908-157963-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0005.flac", "task_type": "lexical_counting", "query": "like", "prompt": "Process the full audio offline. For every occurrence of the word \"like\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.26, 0.55, "one"], [1.91, 2.31, "two"], [3.91, 4.23, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.26,\"end\":0.55},{\"word\":\"two\",\"start\":1.91,\"end\":2.31},{\"word\":\"three\",\"start\":3.91,\"end\":4.23}]"} {"utterance_id": "4077-13751-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0006.flac", "task_type": "lexical_detection", "query": "established", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"established\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.98, 12.68, "established"]], "response_mode": "offline", "answer": "[{\"word\":\"established\",\"start\":11.98,\"end\":12.68}]"} {"utterance_id": "2094-142345-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0037.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.97, 1.38, "servants"]], "response_mode": "offline", "answer": "[{\"word\":\"servants\",\"start\":0.97,\"end\":1.38}]"} {"utterance_id": "8455-210777-0065", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0065.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.46, 3.48, "sir ferdinando"]], "response_mode": "offline", "answer": "[{\"word\":\"sir ferdinando\",\"start\":2.46,\"end\":3.48}]"} {"utterance_id": "7127-75947-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0034.flac", "task_type": "lexical_detection", "query": "consent", "prompt": "After listening to the whole recording, output every instance of the word \"consent\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.73, 2.16, "consent"]], "response_mode": "offline", "answer": "[{\"word\":\"consent\",\"start\":1.73,\"end\":2.16}]"} {"utterance_id": "7176-92135-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0028.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Once the audio has finished, identify every instance of an ordinal number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.23, 2.71, "third"]], "response_mode": "offline", "answer": "[{\"word\":\"third\",\"start\":2.23,\"end\":2.71}]"} {"utterance_id": "7176-88083-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0018.flac", "task_type": "lexical_detection", "query": "valor", "prompt": "Once the audio has finished, identify every instance of the word \"valor\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.64, 4.1, "valor"]], "response_mode": "offline", "answer": "[{\"word\":\"valor\",\"start\":3.64,\"end\":4.1}]"} {"utterance_id": "6829-68771-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0008.flac", "task_type": "lexical_detection", "query": "promised", "prompt": "After the audio ends, extract all occurrences of the word \"promised\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.26, 4.81, "promised"]], "response_mode": "offline", "answer": "[{\"word\":\"promised\",\"start\":4.26,\"end\":4.81}]"} {"utterance_id": "121-123859-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123859/121-123859-0000.flac", "task_type": "lexical_detection", "query": "praises", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"praises\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.69, 7.26, "praises"]], "response_mode": "offline", "answer": "[{\"word\":\"praises\",\"start\":6.69,\"end\":7.26}]"} {"utterance_id": "8463-294828-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0035.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.19, 2.52, "quarter"]], "response_mode": "offline", "answer": "[{\"word\":\"quarter\",\"start\":2.19,\"end\":2.52}]"} {"utterance_id": "5105-28233-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28233/5105-28233-0010.flac", "task_type": "lexical_detection", "query": "stock", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"stock\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.85, 10.26, "stock"]], "response_mode": "offline", "answer": "[{\"word\":\"stock\",\"start\":9.85,\"end\":10.26}]"} {"utterance_id": "5639-40744-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0038.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.04, 9.7, "leocadia"]], "response_mode": "offline", "answer": "[{\"word\":\"leocadia\",\"start\":9.04,\"end\":9.7}]"} {"utterance_id": "8555-284447-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0014.flac", "task_type": "lexical_detection", "query": "goat", "prompt": "After the audio ends, extract all occurrences of the word \"goat\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.33, 2.7, "goat"]], "response_mode": "offline", "answer": "[{\"word\":\"goat\",\"start\":2.33,\"end\":2.7}]"} {"utterance_id": "8230-279154-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0011.flac", "task_type": "lexical_detection", "query": "familiar", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"familiar\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.64, 4.39, "familiar"]], "response_mode": "offline", "answer": "[{\"word\":\"familiar\",\"start\":3.64,\"end\":4.39}]"} {"utterance_id": "3570-5694-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0000.flac", "task_type": "lexical_detection", "query": "emergence", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"emergence\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.77, 4.33, "emergence"]], "response_mode": "offline", "answer": "[{\"word\":\"emergence\",\"start\":3.77,\"end\":4.33}]"} {"utterance_id": "6930-76324-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0005.flac", "task_type": "lexical_detection", "query": "picture", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"picture\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.69, 4.1, "picture"]], "response_mode": "offline", "answer": "[{\"word\":\"picture\",\"start\":3.69,\"end\":4.1}]"} {"utterance_id": "1089-134686-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0023.flac", "task_type": "lexical_detection", "query": "species", "prompt": "Once the audio has finished, identify every instance of the word \"species\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.83, 4.4, "species"]], "response_mode": "offline", "answer": "[{\"word\":\"species\",\"start\":3.83,\"end\":4.4}]"} {"utterance_id": "5683-32879-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0008.flac", "task_type": "lexical_detection", "query": "ill", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"ill\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.39, 1.7, "ill"]], "response_mode": "offline", "answer": "[{\"word\":\"ill\",\"start\":1.39,\"end\":1.7}]"} {"utterance_id": "260-123286-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0010.flac", "task_type": "lexical_detection", "query": "sunday", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"sunday\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 1.12, "sunday"]], "response_mode": "offline", "answer": "[{\"word\":\"sunday\",\"start\":0.52,\"end\":1.12}]"} {"utterance_id": "2961-960-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0004.flac", "task_type": "lexical_detection", "query": "falling", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"falling\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.57, 5.02, "falling"]], "response_mode": "offline", "answer": "[{\"word\":\"falling\",\"start\":4.57,\"end\":5.02}]"} {"utterance_id": "4992-41797-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0010.flac", "task_type": "lexical_detection", "query": "indifferent", "prompt": "After the audio ends, extract all occurrences of the word \"indifferent\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.17, 2.9, "indifferent"]], "response_mode": "offline", "answer": "[{\"word\":\"indifferent\",\"start\":2.17,\"end\":2.9}]"} {"utterance_id": "5142-33396-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0045.flac", "task_type": "lexical_detection", "query": "feast", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"feast\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.03, 1.29, "feast"]], "response_mode": "offline", "answer": "[{\"word\":\"feast\",\"start\":1.03,\"end\":1.29}]"} {"utterance_id": "8224-274381-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0016.flac", "task_type": "lexical_detection", "query": "attained", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"attained\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.94, 10.68, "attained"]], "response_mode": "offline", "answer": "[{\"word\":\"attained\",\"start\":9.94,\"end\":10.68}]"} {"utterance_id": "5683-32879-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0006.flac", "task_type": "lexical_detection", "query": "sweetness", "prompt": "After the audio ends, extract all occurrences of the word \"sweetness\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.16, 1.84, "sweetness"]], "response_mode": "offline", "answer": "[{\"word\":\"sweetness\",\"start\":1.16,\"end\":1.84}]"} {"utterance_id": "672-122797-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0025.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Process the full audio offline. Find every instance of a plant name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.61, 0.9, "tree"], [4.95, 5.37, "trees"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":0.61,\"end\":0.9},{\"word\":\"trees\",\"start\":4.95,\"end\":5.37}]"} {"utterance_id": "7176-92135-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0027.flac", "task_type": "lexical_detection", "query": "larkspur", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"larkspur\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.54, 1.06, "larkspur"]], "response_mode": "offline", "answer": "[{\"word\":\"larkspur\",\"start\":0.54,\"end\":1.06}]"} {"utterance_id": "5683-32879-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0014.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.86, 3.35, "rachel"]], "response_mode": "offline", "answer": "[{\"word\":\"rachel\",\"start\":2.86,\"end\":3.35}]"} {"utterance_id": "672-122797-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0030.flac", "task_type": "lexical_detection", "query": "forest", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"forest\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.27, 2.76, "forest"]], "response_mode": "offline", "answer": "[{\"word\":\"forest\",\"start\":2.27,\"end\":2.76}]"} {"utterance_id": "5683-32865-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0012.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[12.25, 12.76, "captain"]], "response_mode": "offline", "answer": "[{\"word\":\"captain\",\"start\":12.25,\"end\":12.76}]"} {"utterance_id": "8455-210777-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0000.flac", "task_type": "lexical_detection", "query": "acknowledge", "prompt": "Once the audio has finished, identify every instance of the word \"acknowledge\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.13, 3.58, "acknowledge"]], "response_mode": "offline", "answer": "[{\"word\":\"acknowledge\",\"start\":3.13,\"end\":3.58}]"} {"utterance_id": "1320-122617-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0003.flac", "task_type": "lexical_detection", "query": "confusion", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"confusion\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.03, 4.71, "confusion"]], "response_mode": "offline", "answer": "[{\"word\":\"confusion\",\"start\":4.03,\"end\":4.71}]"} {"utterance_id": "7127-75947-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0035.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.4, 1.71, "king"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":1.4,\"end\":1.71}]"} {"utterance_id": "61-70968-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0030.flac", "task_type": "lexical_detection", "query": "shout", "prompt": "When the utterance is complete, report all instances of the word \"shout\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.17, 5.52, "shout"]], "response_mode": "offline", "answer": "[{\"word\":\"shout\",\"start\":5.17,\"end\":5.52}]"} {"utterance_id": "6829-68771-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0018.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.19, 3.5, "maid"]], "response_mode": "offline", "answer": "[{\"word\":\"maid\",\"start\":3.19,\"end\":3.5}]"} {"utterance_id": "2094-142345-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0045.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.05, 2.78, "missus poyser"]], "response_mode": "offline", "answer": "[{\"word\":\"missus poyser\",\"start\":2.05,\"end\":2.78}]"} {"utterance_id": "1089-134691-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0015.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After the audio ends, extract all occurrences of a color name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.48, 3.0, "colours"]], "response_mode": "offline", "answer": "[{\"word\":\"colours\",\"start\":2.48,\"end\":3.0}]"} {"utterance_id": "1284-1181-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0009.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.62, 3.87, "four"]], "response_mode": "offline", "answer": "[{\"word\":\"four\",\"start\":3.62,\"end\":3.87}]"} {"utterance_id": "1089-134686-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0037.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Process the full audio offline. Find every instance of a color name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.61, 1.92, "dark"], [4.12, 4.55, "tawny"]], "response_mode": "offline", "answer": "[{\"word\":\"dark\",\"start\":1.61,\"end\":1.92},{\"word\":\"tawny\",\"start\":4.12,\"end\":4.55}]"} {"utterance_id": "2094-142345-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0008.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Do not answer until the audio ends. Then list each instance of a food name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[28.33, 28.75, "drink"]], "response_mode": "offline", "answer": "[{\"word\":\"drink\",\"start\":28.33,\"end\":28.75}]"} {"utterance_id": "1995-1837-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0001.flac", "task_type": "lexical_detection", "query": "built", "prompt": "After listening to the whole recording, output every instance of the word \"built\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.73, 8.0, "built"]], "response_mode": "offline", "answer": "[{\"word\":\"built\",\"start\":7.73,\"end\":8.0}]"} {"utterance_id": "4077-13754-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0006.flac", "task_type": "lexical_counting", "query": "marriage", "prompt": "After the audio ends, tally every instance of the word \"marriage\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[2.23, 2.63, "one"], [8.35, 8.76, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.23,\"end\":2.63},{\"word\":\"two\",\"start\":8.35,\"end\":8.76}]"} {"utterance_id": "2300-131720-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0021.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Once the audio has finished, identify every instance of an organization name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.23, 6.74, "company"]], "response_mode": "offline", "answer": "[{\"word\":\"company\",\"start\":6.23,\"end\":6.74}]"} {"utterance_id": "5683-32866-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0001.flac", "task_type": "lexical_detection", "query": "less", "prompt": "Once the audio has finished, identify every instance of the word \"less\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.84, 2.16, "less"]], "response_mode": "offline", "answer": "[{\"word\":\"less\",\"start\":1.84,\"end\":2.16}]"} {"utterance_id": "4992-41806-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0002.flac", "task_type": "lexical_counting", "query": "missus", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"missus\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[5.92, 6.28, "one"], [9.59, 9.92, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.92,\"end\":6.28},{\"word\":\"two\",\"start\":9.59,\"end\":9.92}]"} {"utterance_id": "5142-33396-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0045.flac", "task_type": "semantic_detection", "query": "food", "prompt": "After the audio ends, extract all occurrences of a food name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.03, 1.29, "feast"]], "response_mode": "offline", "answer": "[{\"word\":\"feast\",\"start\":1.03,\"end\":1.29}]"} {"utterance_id": "260-123286-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0020.flac", "task_type": "lexical_detection", "query": "august", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"august\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.44, 1.88, "august"]], "response_mode": "offline", "answer": "[{\"word\":\"august\",\"start\":1.44,\"end\":1.88}]"} {"utterance_id": "2094-142345-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0029.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.16, 6.37, "eye"]], "response_mode": "offline", "answer": "[{\"word\":\"eye\",\"start\":6.16,\"end\":6.37}]"} {"utterance_id": "6829-68769-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0030.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.55, 1.28, "bookkeeper"]], "response_mode": "offline", "answer": "[{\"word\":\"bookkeeper\",\"start\":0.55,\"end\":1.28}]"} {"utterance_id": "237-134493-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0018.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After the audio ends, extract all occurrences of a plant name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.67, 4.64, "walnut trees"]], "response_mode": "offline", "answer": "[{\"word\":\"walnut trees\",\"start\":3.67,\"end\":4.64}]"} {"utterance_id": "2830-3980-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0015.flac", "task_type": "lexical_detection", "query": "possess", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"possess\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.98, 1.37, "possess"]], "response_mode": "offline", "answer": "[{\"word\":\"possess\",\"start\":0.98,\"end\":1.37}]"} {"utterance_id": "61-70968-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0024.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.63, 1.08, "squire"]], "response_mode": "offline", "answer": "[{\"word\":\"squire\",\"start\":0.63,\"end\":1.08}]"} {"utterance_id": "8230-279154-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0038.flac", "task_type": "lexical_detection", "query": "nevertheless", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"nevertheless\" with fields word, start, end (seconds). If none, reply [].", "labels": [[16.38, 17.19, "nevertheless"]], "response_mode": "offline", "answer": "[{\"word\":\"nevertheless\",\"start\":16.38,\"end\":17.19}]"} {"utterance_id": "5639-40744-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0023.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.26, 8.59, "mother"], [9.71, 10.1, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":8.26,\"end\":8.59},{\"word\":\"father\",\"start\":9.71,\"end\":10.1}]"} {"utterance_id": "5142-33396-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0037.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.16, 0.64, "hakon"]], "response_mode": "offline", "answer": "[{\"word\":\"hakon\",\"start\":0.16,\"end\":0.64}]"} {"utterance_id": "4507-16021-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0004.flac", "task_type": "lexical_detection", "query": "slang", "prompt": "When the utterance is complete, report all instances of the word \"slang\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.13, 1.76, "slang"]], "response_mode": "offline", "answer": "[{\"word\":\"slang\",\"start\":1.13,\"end\":1.76}]"} {"utterance_id": "5142-36377-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0014.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Once the audio has finished, identify every instance of a country name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[20.63, 21.3, "america"]], "response_mode": "offline", "answer": "[{\"word\":\"america\",\"start\":20.63,\"end\":21.3}]"} {"utterance_id": "61-70968-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0007.flac", "task_type": "lexical_detection", "query": "sister", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"sister\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.41, 1.01, "sister"]], "response_mode": "offline", "answer": "[{\"word\":\"sister\",\"start\":0.41,\"end\":1.01}]"} {"utterance_id": "2961-961-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0018.flac", "task_type": "semantic_detection", "query": "city", "prompt": "You will hear a complete utterance. When it ends, list every instance of a city name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.1, 4.98, "atlantis"]], "response_mode": "offline", "answer": "[{\"word\":\"atlantis\",\"start\":4.1,\"end\":4.98}]"} {"utterance_id": "7127-75947-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0022.flac", "task_type": "lexical_counting", "query": "whoever", "prompt": "After the audio ends, tally every instance of the word \"whoever\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[3.14, 3.56, "one"], [5.45, 5.84, "two"], [8.21, 8.65, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.14,\"end\":3.56},{\"word\":\"two\",\"start\":5.45,\"end\":5.84},{\"word\":\"three\",\"start\":8.21,\"end\":8.65}]"} {"utterance_id": "8555-284447-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0015.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.54, 10.97, "legs"], [12.92, 13.48, "whiskers"]], "response_mode": "offline", "answer": "[{\"word\":\"legs\",\"start\":10.54,\"end\":10.97},{\"word\":\"whiskers\",\"start\":12.92,\"end\":13.48}]"} {"utterance_id": "1580-141084-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0024.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.52, 5.83, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":5.52,\"end\":5.83}]"} {"utterance_id": "4992-41806-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 1.57, "natty harmon"]], "response_mode": "offline", "answer": "[{\"word\":\"natty harmon\",\"start\":0.42,\"end\":1.57}]"} {"utterance_id": "1089-134686-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0003.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.69, 1.28, "bertie"]], "response_mode": "offline", "answer": "[{\"word\":\"bertie\",\"start\":0.69,\"end\":1.28}]"} {"utterance_id": "1221-135767-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0004.flac", "task_type": "lexical_detection", "query": "vivid", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"vivid\" with fields word, start, end (seconds). If none, reply [].", "labels": [[5.97, 6.34, "vivid"]], "response_mode": "offline", "answer": "[{\"word\":\"vivid\",\"start\":5.97,\"end\":6.34}]"} {"utterance_id": "5105-28240-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0014.flac", "task_type": "lexical_detection", "query": "mediterranean", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"mediterranean\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.1, 2.94, "mediterranean"]], "response_mode": "offline", "answer": "[{\"word\":\"mediterranean\",\"start\":2.1,\"end\":2.94}]"} {"utterance_id": "1284-1180-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.5, 1.71, "unc"]], "response_mode": "offline", "answer": "[{\"word\":\"unc\",\"start\":1.5,\"end\":1.71}]"} {"utterance_id": "4446-2271-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0011.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.01, 3.59, "engineer"]], "response_mode": "offline", "answer": "[{\"word\":\"engineer\",\"start\":3.01,\"end\":3.59}]"} {"utterance_id": "1995-1837-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0014.flac", "task_type": "semantic_detection", "query": "color", "prompt": "You will hear a complete utterance. When it ends, list every instance of a color name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.1, 9.94, "olive green"]], "response_mode": "offline", "answer": "[{\"word\":\"olive green\",\"start\":9.1,\"end\":9.94}]"} {"utterance_id": "5639-40744-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0016.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[6.52, 6.88, "half"]], "response_mode": "offline", "answer": "[{\"word\":\"half\",\"start\":6.52,\"end\":6.88}]"} {"utterance_id": "1089-134686-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0026.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.65, 1.0, "rector"]], "response_mode": "offline", "answer": "[{\"word\":\"rector\",\"start\":0.65,\"end\":1.0}]"} {"utterance_id": "1320-122617-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0022.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen carefully to the entire clip, then report every instance of an organization name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.3, 0.88, "delawares"]], "response_mode": "offline", "answer": "[{\"word\":\"delawares\",\"start\":0.3,\"end\":0.88}]"} {"utterance_id": "4970-29095-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0030.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the full audio. After it finishes, find every instance of a family relation. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.49, 1.14, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":0.49,\"end\":1.14}]"} {"utterance_id": "61-70968-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0022.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.04, 4.54, "apprentice"]], "response_mode": "offline", "answer": "[{\"word\":\"apprentice\",\"start\":4.04,\"end\":4.54}]"} {"utterance_id": "7176-92135-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0038.flac", "task_type": "semantic_detection", "query": "food", "prompt": "After the audio ends, extract all occurrences of a food name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.68, 0.91, "meal"]], "response_mode": "offline", "answer": "[{\"word\":\"meal\",\"start\":0.68,\"end\":0.91}]"} {"utterance_id": "4446-2273-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.36, 10.7, "bartley"]], "response_mode": "offline", "answer": "[{\"word\":\"bartley\",\"start\":10.36,\"end\":10.7}]"} {"utterance_id": "4446-2275-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0042.flac", "task_type": "lexical_detection", "query": "difference", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"difference\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.27, 4.94, "difference"]], "response_mode": "offline", "answer": "[{\"word\":\"difference\",\"start\":4.27,\"end\":4.94}]"} {"utterance_id": "2830-3979-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.46, 1.86, "luther's"]], "response_mode": "offline", "answer": "[{\"word\":\"luther's\",\"start\":1.46,\"end\":1.86}]"} {"utterance_id": "260-123288-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0015.flac", "task_type": "lexical_counting", "query": "continual", "prompt": "Process the full audio offline. For every occurrence of the word \"continual\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.7, 3.26, "one"], [6.33, 6.91, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.7,\"end\":3.26},{\"word\":\"two\",\"start\":6.33,\"end\":6.91}]"} {"utterance_id": "61-70970-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0034.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.5, 3.88, "half"]], "response_mode": "offline", "answer": "[{\"word\":\"half\",\"start\":3.5,\"end\":3.88}]"} {"utterance_id": "5142-33396-0068", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0068.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.97, 2.37, "thrall"]], "response_mode": "offline", "answer": "[{\"word\":\"thrall\",\"start\":1.97,\"end\":2.37}]"} {"utterance_id": "4077-13754-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0013.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Do not answer until the audio ends. Then list each instance of an organization name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.34, 4.68, "court"]], "response_mode": "offline", "answer": "[{\"word\":\"court\",\"start\":4.34,\"end\":4.68}]"} {"utterance_id": "672-122797-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0020.flac", "task_type": "semantic_detection", "query": "color", "prompt": "You will hear a complete utterance. When it ends, list every instance of a color name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.41, 6.93, "green"]], "response_mode": "offline", "answer": "[{\"word\":\"green\",\"start\":6.41,\"end\":6.93}]"} {"utterance_id": "2300-131720-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.38, 14.85, "edison"]], "response_mode": "offline", "answer": "[{\"word\":\"edison\",\"start\":14.38,\"end\":14.85}]"} {"utterance_id": "2094-142345-0057", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0057.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Process the full audio offline. Find every instance of a country name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.8, 9.18, "kingdom"]], "response_mode": "offline", "answer": "[{\"word\":\"kingdom\",\"start\":8.8,\"end\":9.18}]"} {"utterance_id": "8455-210777-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0022.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.9, 1.66, "missus neverbend"]], "response_mode": "offline", "answer": "[{\"word\":\"missus neverbend\",\"start\":0.9,\"end\":1.66}]"} {"utterance_id": "2094-142345-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0029.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.87, 3.66, "missus poyser"]], "response_mode": "offline", "answer": "[{\"word\":\"missus poyser\",\"start\":2.87,\"end\":3.66}]"} {"utterance_id": "2830-3980-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0043.flac", "task_type": "lexical_detection", "query": "proves", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"proves\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.01, 1.33, "proves"]], "response_mode": "offline", "answer": "[{\"word\":\"proves\",\"start\":1.01,\"end\":1.33}]"} {"utterance_id": "1320-122617-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0026.flac", "task_type": "lexical_counting", "query": "done", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"done\" with word/start/end. If none, reply [].", "labels": [[1.57, 1.9, "one"], [3.61, 3.84, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.57,\"end\":1.9},{\"word\":\"two\",\"start\":3.61,\"end\":3.84}]"} {"utterance_id": "1284-1180-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0000.flac", "task_type": "lexical_detection", "query": "wore", "prompt": "Once the audio has finished, identify every instance of the word \"wore\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.37, 0.63, "wore"]], "response_mode": "offline", "answer": "[{\"word\":\"wore\",\"start\":0.37,\"end\":0.63}]"} {"utterance_id": "6930-75918-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0008.flac", "task_type": "lexical_detection", "query": "suspect", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"suspect\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.7, 4.46, "suspect"]], "response_mode": "offline", "answer": "[{\"word\":\"suspect\",\"start\":3.7,\"end\":4.46}]"} {"utterance_id": "4507-16021-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0047.flac", "task_type": "lexical_detection", "query": "affairs", "prompt": "When the utterance is complete, report all instances of the word \"affairs\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[33.91, 34.69, "affairs"]], "response_mode": "offline", "answer": "[{\"word\":\"affairs\",\"start\":33.91,\"end\":34.69}]"} {"utterance_id": "7176-88083-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0017.flac", "task_type": "lexical_detection", "query": "rapids", "prompt": "Once the audio has finished, identify every instance of the word \"rapids\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.1, 1.49, "rapids"]], "response_mode": "offline", "answer": "[{\"word\":\"rapids\",\"start\":1.1,\"end\":1.49}]"} {"utterance_id": "1580-141084-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0021.flac", "task_type": "lexical_detection", "query": "pyramids", "prompt": "Once the audio has finished, identify every instance of the word \"pyramids\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.4, 1.9, "pyramids"]], "response_mode": "offline", "answer": "[{\"word\":\"pyramids\",\"start\":1.4,\"end\":1.9}]"} {"utterance_id": "1284-1180-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0032.flac", "task_type": "lexical_detection", "query": "cupboard", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"cupboard\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.61, 4.05, "cupboard"]], "response_mode": "offline", "answer": "[{\"word\":\"cupboard\",\"start\":3.61,\"end\":4.05}]"} {"utterance_id": "5639-40744-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0022.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[8.82, 9.44, "midwife"]], "response_mode": "offline", "answer": "[{\"word\":\"midwife\",\"start\":8.82,\"end\":9.44}]"} {"utterance_id": "7176-92135-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0012.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.31, 7.61, "thirty"]], "response_mode": "offline", "answer": "[{\"word\":\"thirty\",\"start\":7.31,\"end\":7.61}]"} {"utterance_id": "237-134500-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0011.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.64, 2.02, "spider"]], "response_mode": "offline", "answer": "[{\"word\":\"spider\",\"start\":1.64,\"end\":2.02}]"} {"utterance_id": "61-70970-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0022.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.54, 3.06, "esquire"]], "response_mode": "offline", "answer": "[{\"word\":\"esquire\",\"start\":2.54,\"end\":3.06}]"} {"utterance_id": "1580-141083-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0016.flac", "task_type": "lexical_detection", "query": "hurry", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"hurry\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.91, 1.14, "hurry"]], "response_mode": "offline", "answer": "[{\"word\":\"hurry\",\"start\":0.91,\"end\":1.14}]"} {"utterance_id": "4507-16021-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0037.flac", "task_type": "lexical_detection", "query": "clothes", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"clothes\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.69, 1.49, "clothes"]], "response_mode": "offline", "answer": "[{\"word\":\"clothes\",\"start\":0.69,\"end\":1.49}]"} {"utterance_id": "672-122797-0063", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0063.flac", "task_type": "lexical_detection", "query": "went", "prompt": "Process the full audio offline. Find every instance of the word \"went\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.62, 4.07, "went"]], "response_mode": "offline", "answer": "[{\"word\":\"went\",\"start\":3.62,\"end\":4.07}]"} {"utterance_id": "1580-141084-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0047.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "After the audio ends, extract all occurrences of a time of day. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.2, 3.59, "morning"]], "response_mode": "offline", "answer": "[{\"word\":\"morning\",\"start\":3.2,\"end\":3.59}]"} {"utterance_id": "6829-68771-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0018.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.51, 7.99, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":7.51,\"end\":7.99}]"} {"utterance_id": "8455-210777-0056", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0056.flac", "task_type": "lexical_detection", "query": "colony", "prompt": "Process the full audio offline. Find every instance of the word \"colony\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.83, 3.27, "colony"]], "response_mode": "offline", "answer": "[{\"word\":\"colony\",\"start\":2.83,\"end\":3.27}]"} {"utterance_id": "237-134493-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0017.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[2.93, 3.09, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.93,\"end\":3.09}]"} {"utterance_id": "7729-102255-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0041.flac", "task_type": "lexical_detection", "query": "childish", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"childish\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.4, 2.11, "childish"]], "response_mode": "offline", "answer": "[{\"word\":\"childish\",\"start\":1.4,\"end\":2.11}]"} {"utterance_id": "4970-29095-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0021.flac", "task_type": "lexical_detection", "query": "ghosts", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"ghosts\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.29, 8.71, "ghosts"]], "response_mode": "offline", "answer": "[{\"word\":\"ghosts\",\"start\":8.29,\"end\":8.71}]"} {"utterance_id": "61-70968-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0024.flac", "task_type": "lexical_detection", "query": "squire", "prompt": "Process the full audio offline. Find every instance of the word \"squire\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.63, 1.08, "squire"]], "response_mode": "offline", "answer": "[{\"word\":\"squire\",\"start\":0.63,\"end\":1.08}]"} {"utterance_id": "61-70968-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0003.flac", "task_type": "lexical_counting", "query": "father", "prompt": "Listen to the full audio. Count every occurrence of the word \"father\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.15, 1.66, "one"], [3.47, 3.98, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.15,\"end\":1.66},{\"word\":\"two\",\"start\":3.47,\"end\":3.98}]"} {"utterance_id": "2830-3980-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0005.flac", "task_type": "lexical_detection", "query": "lutheran", "prompt": "When the utterance is complete, report all instances of the word \"lutheran\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.74, 3.15, "lutheran"]], "response_mode": "offline", "answer": "[{\"word\":\"lutheran\",\"start\":2.74,\"end\":3.15}]"} {"utterance_id": "4970-29095-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0016.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 0.91, "ruth"]], "response_mode": "offline", "answer": "[{\"word\":\"ruth\",\"start\":0.45,\"end\":0.91}]"} {"utterance_id": "7729-102255-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0003.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[10.83, 11.53, "two thousand"]], "response_mode": "offline", "answer": "[{\"word\":\"two thousand\",\"start\":10.83,\"end\":11.53}]"} {"utterance_id": "5142-33396-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0041.flac", "task_type": "lexical_detection", "query": "guards", "prompt": "After the audio ends, extract all occurrences of the word \"guards\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.71, 1.12, "guards"]], "response_mode": "offline", "answer": "[{\"word\":\"guards\",\"start\":0.71,\"end\":1.12}]"} {"utterance_id": "6829-68771-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0004.flac", "task_type": "lexical_detection", "query": "reynolds", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"reynolds\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.23, 2.74, "reynolds"]], "response_mode": "offline", "answer": "[{\"word\":\"reynolds\",\"start\":2.23,\"end\":2.74}]"} {"utterance_id": "2300-131720-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0035.flac", "task_type": "lexical_detection", "query": "association", "prompt": "After listening to the whole recording, output every instance of the word \"association\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.1, 3.96, "association"]], "response_mode": "offline", "answer": "[{\"word\":\"association\",\"start\":3.1,\"end\":3.96}]"} {"utterance_id": "3570-5695-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0005.flac", "task_type": "lexical_detection", "query": "descent", "prompt": "Process the full audio offline. Find every instance of the word \"descent\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.72, 6.19, "descent"]], "response_mode": "offline", "answer": "[{\"word\":\"descent\",\"start\":5.72,\"end\":6.19}]"} {"utterance_id": "3570-5694-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0007.flac", "task_type": "lexical_detection", "query": "conditions", "prompt": "Once the audio has finished, identify every instance of the word \"conditions\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.2, 3.78, "conditions"]], "response_mode": "offline", "answer": "[{\"word\":\"conditions\",\"start\":3.2,\"end\":3.78}]"} {"utterance_id": "5683-32865-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0005.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.26, 2.07, "captain lake"]], "response_mode": "offline", "answer": "[{\"word\":\"captain lake\",\"start\":1.26,\"end\":2.07}]"} {"utterance_id": "5639-40744-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0002.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.7, 8.5, "servant maid"]], "response_mode": "offline", "answer": "[{\"word\":\"servant maid\",\"start\":7.7,\"end\":8.5}]"} {"utterance_id": "672-122797-0056", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0056.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Process the full audio offline. Find every instance of a plant name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.69, 1.56, "fir tree"]], "response_mode": "offline", "answer": "[{\"word\":\"fir tree\",\"start\":0.69,\"end\":1.56}]"} {"utterance_id": "4446-2273-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0016.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.69, 5.91, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":5.69,\"end\":5.91}]"} {"utterance_id": "1284-1181-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0005.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.95, 2.65, "brains"]], "response_mode": "offline", "answer": "[{\"word\":\"brains\",\"start\":1.95,\"end\":2.65}]"} {"utterance_id": "4507-16021-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0040.flac", "task_type": "lexical_detection", "query": "hears", "prompt": "After listening to the whole recording, output every instance of the word \"hears\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.61, 2.07, "hears"]], "response_mode": "offline", "answer": "[{\"word\":\"hears\",\"start\":1.61,\"end\":2.07}]"} {"utterance_id": "8463-287645-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0007.flac", "task_type": "lexical_counting", "query": "mistress", "prompt": "After the audio ends, tally every instance of the word \"mistress\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[11.33, 12.01, "one"], [13.18, 13.59, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":11.33,\"end\":12.01},{\"word\":\"two\",\"start\":13.18,\"end\":13.59}]"} {"utterance_id": "5639-40744-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0022.flac", "task_type": "semantic_counting", "query": "kinship", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a family relation: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[7.02, 7.34, "one"], [10.8, 11.22, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.02,\"end\":7.34},{\"word\":\"two\",\"start\":10.8,\"end\":11.22}]"} {"utterance_id": "908-31957-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0003.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a plant name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.78, 4.47, "flowers"]], "response_mode": "offline", "answer": "[{\"word\":\"flowers\",\"start\":3.78,\"end\":4.47}]"} {"utterance_id": "7021-79730-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0004.flac", "task_type": "lexical_detection", "query": "countenance", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"countenance\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.22, 7.96, "countenance"]], "response_mode": "offline", "answer": "[{\"word\":\"countenance\",\"start\":7.22,\"end\":7.96}]"} {"utterance_id": "1995-1837-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0013.flac", "task_type": "lexical_detection", "query": "lagoon", "prompt": "Process the full audio offline. Find every instance of the word \"lagoon\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.88, 2.28, "lagoon"]], "response_mode": "offline", "answer": "[{\"word\":\"lagoon\",\"start\":1.88,\"end\":2.28}]"} {"utterance_id": "4992-23283-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0005.flac", "task_type": "lexical_detection", "query": "perfect", "prompt": "After listening to the whole recording, output every instance of the word \"perfect\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.79, 8.26, "perfect"]], "response_mode": "offline", "answer": "[{\"word\":\"perfect\",\"start\":7.79,\"end\":8.26}]"} {"utterance_id": "4077-13751-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0012.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.49, 1.48, "lieutenant governor"], [3.86, 4.39, "governor"]], "response_mode": "offline", "answer": "[{\"word\":\"lieutenant governor\",\"start\":0.49,\"end\":1.48},{\"word\":\"governor\",\"start\":3.86,\"end\":4.39}]"} {"utterance_id": "7176-92135-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0014.flac", "task_type": "lexical_counting", "query": "soliloquy", "prompt": "Process the full audio offline. For every occurrence of the word \"soliloquy\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.56, 7.38, "one"], [13.03, 13.7, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.56,\"end\":7.38},{\"word\":\"two\",\"start\":13.03,\"end\":13.7}]"} {"utterance_id": "672-122797-0057", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0057.flac", "task_type": "lexical_detection", "query": "reality", "prompt": "After the audio ends, extract all occurrences of the word \"reality\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.11, 3.0, "reality"]], "response_mode": "offline", "answer": "[{\"word\":\"reality\",\"start\":2.11,\"end\":3.0}]"} {"utterance_id": "6829-68769-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0048.flac", "task_type": "lexical_counting", "query": "check", "prompt": "Once the audio has finished, list each occurrence of the word \"check\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[0.68, 0.99, "one"], [4.19, 4.51, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.68,\"end\":0.99},{\"word\":\"two\",\"start\":4.19,\"end\":4.51}]"} {"utterance_id": "4992-23283-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0018.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.17, 2.51, "hand"], [3.62, 3.93, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":2.17,\"end\":2.51},{\"word\":\"heart\",\"start\":3.62,\"end\":3.93}]"} {"utterance_id": "1580-141084-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0038.flac", "task_type": "lexical_detection", "query": "chance", "prompt": "When the utterance is complete, report all instances of the word \"chance\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.12, 2.74, "chance"]], "response_mode": "offline", "answer": "[{\"word\":\"chance\",\"start\":2.12,\"end\":2.74}]"} {"utterance_id": "8463-294828-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0017.flac", "task_type": "lexical_detection", "query": "trunk", "prompt": "After the audio ends, extract all occurrences of the word \"trunk\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.24, 1.62, "trunk"]], "response_mode": "offline", "answer": "[{\"word\":\"trunk\",\"start\":1.24,\"end\":1.62}]"} {"utterance_id": "1284-1180-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0014.flac", "task_type": "lexical_detection", "query": "meal", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"meal\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.38, 2.68, "meal"]], "response_mode": "offline", "answer": "[{\"word\":\"meal\",\"start\":2.38,\"end\":2.68}]"} {"utterance_id": "1188-133604-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0035.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.39, 1.87, "chaucer's"]], "response_mode": "offline", "answer": "[{\"word\":\"chaucer's\",\"start\":1.39,\"end\":1.87}]"} {"utterance_id": "908-31957-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0025.flac", "task_type": "lexical_counting", "query": "thee", "prompt": "Process the full audio offline. For every occurrence of the word \"thee\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.0, 1.27, "one"], [6.69, 7.01, "two"], [18.08, 18.24, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.0,\"end\":1.27},{\"word\":\"two\",\"start\":6.69,\"end\":7.01},{\"word\":\"three\",\"start\":18.08,\"end\":18.24}]"} {"utterance_id": "672-122797-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0021.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an ordinal number with fields word, start, end (seconds). If none, reply [].", "labels": [[2.19, 2.66, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":2.19,\"end\":2.66}]"} {"utterance_id": "4507-16021-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0035.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.35, 5.12, "historian"]], "response_mode": "offline", "answer": "[{\"word\":\"historian\",\"start\":4.35,\"end\":5.12}]"} {"utterance_id": "4970-29093-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0002.flac", "task_type": "lexical_detection", "query": "occupation", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"occupation\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.96, 6.68, "occupation"]], "response_mode": "offline", "answer": "[{\"word\":\"occupation\",\"start\":5.96,\"end\":6.68}]"} {"utterance_id": "237-134500-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0041.flac", "task_type": "lexical_counting", "query": "pray", "prompt": "When the utterance is complete, report how the word \"pray\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.15, 1.51, "one"], [5.38, 5.66, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.15,\"end\":1.51},{\"word\":\"two\",\"start\":5.38,\"end\":5.66}]"} {"utterance_id": "5142-36377-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0007.flac", "task_type": "semantic_detection", "query": "food", "prompt": "After the audio ends, extract all occurrences of a food name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.31, 5.7, "supper"]], "response_mode": "offline", "answer": "[{\"word\":\"supper\",\"start\":5.31,\"end\":5.7}]"} {"utterance_id": "4992-41806-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0014.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[10.24, 10.46, "one"], [12.68, 13.41, "eleven"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":10.24,\"end\":10.46},{\"word\":\"eleven\",\"start\":12.68,\"end\":13.41}]"} {"utterance_id": "3575-170457-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0040.flac", "task_type": "lexical_detection", "query": "admitted", "prompt": "Process the full audio offline. Find every instance of the word \"admitted\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.25, 3.73, "admitted"]], "response_mode": "offline", "answer": "[{\"word\":\"admitted\",\"start\":3.25,\"end\":3.73}]"} {"utterance_id": "2094-142345-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0043.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen to the full audio. After it finishes, find every instance of an animal name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.92, 4.59, "horses"]], "response_mode": "offline", "answer": "[{\"word\":\"horses\",\"start\":3.92,\"end\":4.59}]"} {"utterance_id": "2830-3980-0071", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0071.flac", "task_type": "lexical_detection", "query": "merit", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"merit\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.07, 2.48, "merit"]], "response_mode": "offline", "answer": "[{\"word\":\"merit\",\"start\":2.07,\"end\":2.48}]"} {"utterance_id": "2961-961-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0005.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.28, 1.82, "solon"]], "response_mode": "offline", "answer": "[{\"word\":\"solon\",\"start\":1.28,\"end\":1.82}]"} {"utterance_id": "6829-68769-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0014.flac", "task_type": "lexical_detection", "query": "succession", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"succession\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.9, 2.5, "succession"]], "response_mode": "offline", "answer": "[{\"word\":\"succession\",\"start\":1.9,\"end\":2.5}]"} {"utterance_id": "672-122797-0071", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0071.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Once the audio has finished, identify every instance of a plant name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.63, 6.33, "fir tree"]], "response_mode": "offline", "answer": "[{\"word\":\"fir tree\",\"start\":5.63,\"end\":6.33}]"} {"utterance_id": "7127-75946-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0029.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[0.62, 1.08, "king"], [4.8, 5.28, "dancer"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":0.62,\"end\":1.08},{\"word\":\"dancer\",\"start\":4.8,\"end\":5.28}]"} {"utterance_id": "672-122797-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0029.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "You will hear a complete utterance. When it ends, list every instance of a time of day you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.99, 2.53, "evening"]], "response_mode": "offline", "answer": "[{\"word\":\"evening\",\"start\":1.99,\"end\":2.53}]"} {"utterance_id": "1580-141083-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0014.flac", "task_type": "lexical_detection", "query": "rolled", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"rolled\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.91, 2.19, "rolled"]], "response_mode": "offline", "answer": "[{\"word\":\"rolled\",\"start\":1.91,\"end\":2.19}]"} {"utterance_id": "6829-68769-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0008.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.75, 1.04, "twelve"]], "response_mode": "offline", "answer": "[{\"word\":\"twelve\",\"start\":0.75,\"end\":1.04}]"} {"utterance_id": "1320-122617-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0037.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[5.78, 6.27, "features"]], "response_mode": "offline", "answer": "[{\"word\":\"features\",\"start\":5.78,\"end\":6.27}]"} {"utterance_id": "7729-102255-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0030.flac", "task_type": "lexical_detection", "query": "assisted", "prompt": "After listening to the whole recording, output every instance of the word \"assisted\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.32, 6.78, "assisted"]], "response_mode": "offline", "answer": "[{\"word\":\"assisted\",\"start\":6.32,\"end\":6.78}]"} {"utterance_id": "672-122797-0072", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0072.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[0.78, 1.42, "gardener's"]], "response_mode": "offline", "answer": "[{\"word\":\"gardener's\",\"start\":0.78,\"end\":1.42}]"} {"utterance_id": "1320-122612-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0000.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Process the full audio offline. Find every instance of a country name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.43, 3.94, "country"]], "response_mode": "offline", "answer": "[{\"word\":\"country\",\"start\":3.43,\"end\":3.94}]"} {"utterance_id": "260-123288-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0027.flac", "task_type": "lexical_detection", "query": "fills", "prompt": "After listening to the whole recording, output every instance of the word \"fills\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.27, 2.62, "fills"], [5.0, 5.35, "fills"]], "response_mode": "offline", "answer": "[{\"word\":\"fills\",\"start\":2.27,\"end\":2.62},{\"word\":\"fills\",\"start\":5.0,\"end\":5.35}]"} {"utterance_id": "4970-29093-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.5, 0.84, "philip"]], "response_mode": "offline", "answer": "[{\"word\":\"philip\",\"start\":0.5,\"end\":0.84}]"} {"utterance_id": "7127-75946-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0017.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.53, 1.16, "legs"], [5.36, 5.68, "flesh"], [11.14, 11.45, "flesh"]], "response_mode": "offline", "answer": "[{\"word\":\"legs\",\"start\":0.53,\"end\":1.16},{\"word\":\"flesh\",\"start\":5.36,\"end\":5.68},{\"word\":\"flesh\",\"start\":11.14,\"end\":11.45}]"} {"utterance_id": "8455-210777-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0008.flac", "task_type": "lexical_detection", "query": "compliance", "prompt": "After listening to the whole recording, output every instance of the word \"compliance\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.58, 2.23, "compliance"]], "response_mode": "offline", "answer": "[{\"word\":\"compliance\",\"start\":1.58,\"end\":2.23}]"} {"utterance_id": "61-70970-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0027.flac", "task_type": "lexical_detection", "query": "carefully", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"carefully\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.79, 1.27, "carefully"]], "response_mode": "offline", "answer": "[{\"word\":\"carefully\",\"start\":0.79,\"end\":1.27}]"} {"utterance_id": "237-134500-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0019.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.93, 2.31, "lap"]], "response_mode": "offline", "answer": "[{\"word\":\"lap\",\"start\":1.93,\"end\":2.31}]"} {"utterance_id": "5142-36377-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0006.flac", "task_type": "lexical_detection", "query": "member", "prompt": "After listening to the whole recording, output every instance of the word \"member\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.39, 0.68, "member"]], "response_mode": "offline", "answer": "[{\"word\":\"member\",\"start\":0.39,\"end\":0.68}]"} {"utterance_id": "7729-102255-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0032.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an organization name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.54, 1.79, "free state hotel"]], "response_mode": "offline", "answer": "[{\"word\":\"free state hotel\",\"start\":0.54,\"end\":1.79}]"} {"utterance_id": "1320-122617-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.26, 1.62, "uncas"]], "response_mode": "offline", "answer": "[{\"word\":\"uncas\",\"start\":1.26,\"end\":1.62}]"} {"utterance_id": "908-157963-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0030.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.59, 1.11, "ear"]], "response_mode": "offline", "answer": "[{\"word\":\"ear\",\"start\":0.59,\"end\":1.11}]"} {"utterance_id": "5142-33396-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0040.flac", "task_type": "lexical_detection", "query": "same", "prompt": "Once the audio has finished, identify every instance of the word \"same\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.89, 2.26, "same"]], "response_mode": "offline", "answer": "[{\"word\":\"same\",\"start\":1.89,\"end\":2.26}]"} {"utterance_id": "1580-141084-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0009.flac", "task_type": "lexical_detection", "query": "indicate", "prompt": "Process the full audio offline. Find every instance of the word \"indicate\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.85, 3.35, "indicate"]], "response_mode": "offline", "answer": "[{\"word\":\"indicate\",\"start\":2.85,\"end\":3.35}]"} {"utterance_id": "8230-279154-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0026.flac", "task_type": "lexical_detection", "query": "understanding", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"understanding\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.25, 2.97, "understanding"]], "response_mode": "offline", "answer": "[{\"word\":\"understanding\",\"start\":2.25,\"end\":2.97}]"} {"utterance_id": "8555-284449-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0009.flac", "task_type": "lexical_detection", "query": "mate", "prompt": "Process the full audio offline. Find every instance of the word \"mate\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.74, 1.24, "mate"]], "response_mode": "offline", "answer": "[{\"word\":\"mate\",\"start\":0.74,\"end\":1.24}]"} {"utterance_id": "1188-133604-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0032.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.46, 3.7, "edward burne jones"], [5.38, 6.01, "psyche"], [8.11, 8.68, "psyche"]], "response_mode": "offline", "answer": "[{\"word\":\"edward burne jones\",\"start\":2.46,\"end\":3.7},{\"word\":\"psyche\",\"start\":5.38,\"end\":6.01},{\"word\":\"psyche\",\"start\":8.11,\"end\":8.68}]"} {"utterance_id": "1995-1837-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0023.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Do not answer until the audio ends. Then list each instance of an ordinal number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[13.31, 13.85, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":13.31,\"end\":13.85}]"} {"utterance_id": "1221-135766-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0015.flac", "task_type": "lexical_detection", "query": "spoken", "prompt": "Process the full audio offline. Find every instance of the word \"spoken\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 0.82, "spoken"]], "response_mode": "offline", "answer": "[{\"word\":\"spoken\",\"start\":0.42,\"end\":0.82}]"} {"utterance_id": "61-70970-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0009.flac", "task_type": "lexical_detection", "query": "tis", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"tis\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.15, 0.47, "tis"]], "response_mode": "offline", "answer": "[{\"word\":\"tis\",\"start\":0.15,\"end\":0.47}]"} {"utterance_id": "4970-29095-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0006.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After the audio ends, extract all occurrences of an organization name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.62, 3.01, "school"]], "response_mode": "offline", "answer": "[{\"word\":\"school\",\"start\":2.62,\"end\":3.01}]"} {"utterance_id": "4446-2275-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0016.flac", "task_type": "lexical_detection", "query": "breathing", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"breathing\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.2, 3.68, "breathing"]], "response_mode": "offline", "answer": "[{\"word\":\"breathing\",\"start\":3.2,\"end\":3.68}]"} {"utterance_id": "1284-1181-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0007.flac", "task_type": "lexical_detection", "query": "bottles", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"bottles\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.15, 3.76, "bottles"]], "response_mode": "offline", "answer": "[{\"word\":\"bottles\",\"start\":3.15,\"end\":3.76}]"} {"utterance_id": "237-134493-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0010.flac", "task_type": "lexical_detection", "query": "scythe", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"scythe\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.3, 1.79, "scythe"]], "response_mode": "offline", "answer": "[{\"word\":\"scythe\",\"start\":1.3,\"end\":1.79}]"} {"utterance_id": "3570-5695-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0010.flac", "task_type": "lexical_detection", "query": "organization", "prompt": "After listening to the whole recording, output every instance of the word \"organization\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.8, 1.47, "organization"]], "response_mode": "offline", "answer": "[{\"word\":\"organization\",\"start\":0.8,\"end\":1.47}]"} {"utterance_id": "4077-13751-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0014.flac", "task_type": "semantic_counting", "query": "state_or_province", "prompt": "Listen to the entire clip, then report the count sequence for a state or province name as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.67, 6.36, "one"], [8.76, 9.75, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.67,\"end\":6.36},{\"word\":\"two\",\"start\":8.76,\"end\":9.75}]"} {"utterance_id": "2094-142345-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0035.flac", "task_type": "lexical_detection", "query": "aunt", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"aunt\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.93, 2.33, "aunt"]], "response_mode": "offline", "answer": "[{\"word\":\"aunt\",\"start\":1.93,\"end\":2.33}]"} {"utterance_id": "3729-6852-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0025.flac", "task_type": "lexical_detection", "query": "meridian", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"meridian\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.39, 1.94, "meridian"]], "response_mode": "offline", "answer": "[{\"word\":\"meridian\",\"start\":1.39,\"end\":1.94}]"} {"utterance_id": "4507-16021-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0007.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.16, 1.53, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.16,\"end\":1.53}]"} {"utterance_id": "4992-23283-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0012.flac", "task_type": "lexical_detection", "query": "softness", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"softness\" with fields word, start, end (seconds). If none, reply [].", "labels": [[14.05, 14.66, "softness"]], "response_mode": "offline", "answer": "[{\"word\":\"softness\",\"start\":14.05,\"end\":14.66}]"} {"utterance_id": "2830-3980-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0000.flac", "task_type": "lexical_detection", "query": "sought", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"sought\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.35, 1.65, "sought"]], "response_mode": "offline", "answer": "[{\"word\":\"sought\",\"start\":1.35,\"end\":1.65}]"} {"utterance_id": "1580-141083-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0042.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 0.84, "scholar"]], "response_mode": "offline", "answer": "[{\"word\":\"scholar\",\"start\":0.45,\"end\":0.84}]"} {"utterance_id": "7021-85628-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0001.flac", "task_type": "lexical_detection", "query": "near", "prompt": "Once the audio has finished, identify every instance of the word \"near\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.16, 4.48, "near"]], "response_mode": "offline", "answer": "[{\"word\":\"near\",\"start\":4.16,\"end\":4.48}]"} {"utterance_id": "1089-134691-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0012.flac", "task_type": "lexical_counting", "query": "generous", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"generous\" (word, start, end). If none, reply [].", "labels": [[2.67, 3.19, "one"], [11.35, 11.89, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.67,\"end\":3.19},{\"word\":\"two\",\"start\":11.35,\"end\":11.89}]"} {"utterance_id": "1580-141084-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0047.flac", "task_type": "lexical_detection", "query": "middle", "prompt": "After listening to the whole recording, output every instance of the word \"middle\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.75, 3.94, "middle"]], "response_mode": "offline", "answer": "[{\"word\":\"middle\",\"start\":3.75,\"end\":3.94}]"} {"utterance_id": "3729-6852-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0035.flac", "task_type": "lexical_detection", "query": "princes", "prompt": "After listening to the whole recording, output every instance of the word \"princes\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.86, 8.44, "princes"]], "response_mode": "offline", "answer": "[{\"word\":\"princes\",\"start\":7.86,\"end\":8.44}]"} {"utterance_id": "2300-131720-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0030.flac", "task_type": "lexical_detection", "query": "employed", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"employed\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.01, 1.6, "employed"]], "response_mode": "offline", "answer": "[{\"word\":\"employed\",\"start\":1.01,\"end\":1.6}]"} {"utterance_id": "8455-210777-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0031.flac", "task_type": "lexical_detection", "query": "courtesy", "prompt": "After listening to the whole recording, output every instance of the word \"courtesy\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.33, 2.74, "courtesy"]], "response_mode": "offline", "answer": "[{\"word\":\"courtesy\",\"start\":2.33,\"end\":2.74}]"} {"utterance_id": "3729-6852-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0033.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the full audio. After it finishes, find every instance of a family relation. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[26.74, 27.28, "daughter"]], "response_mode": "offline", "answer": "[{\"word\":\"daughter\",\"start\":26.74,\"end\":27.28}]"} {"utterance_id": "8455-210777-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0024.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an ordinal number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.77, 6.12, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":5.77,\"end\":6.12}]"} {"utterance_id": "61-70968-0056", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0056.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.93, 3.4, "cheeks"]], "response_mode": "offline", "answer": "[{\"word\":\"cheeks\",\"start\":2.93,\"end\":3.4}]"} {"utterance_id": "4507-16021-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0010.flac", "task_type": "lexical_detection", "query": "fulfilled", "prompt": "When the utterance is complete, report all instances of the word \"fulfilled\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.5, 14.16, "fulfilled"]], "response_mode": "offline", "answer": "[{\"word\":\"fulfilled\",\"start\":13.5,\"end\":14.16}]"} {"utterance_id": "4077-13751-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0000.flac", "task_type": "lexical_detection", "query": "april", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"april\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.17, 1.61, "april"]], "response_mode": "offline", "answer": "[{\"word\":\"april\",\"start\":1.17,\"end\":1.61}]"} {"utterance_id": "908-31957-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0023.flac", "task_type": "lexical_detection", "query": "strive", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"strive\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.6, 3.33, "strive"]], "response_mode": "offline", "answer": "[{\"word\":\"strive\",\"start\":2.6,\"end\":3.33}]"} {"utterance_id": "260-123288-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0005.flac", "task_type": "semantic_detection", "query": "color", "prompt": "You will hear a complete utterance. When it ends, list every instance of a color name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.51, 8.04, "grey"]], "response_mode": "offline", "answer": "[{\"word\":\"grey\",\"start\":7.51,\"end\":8.04}]"} {"utterance_id": "260-123440-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0008.flac", "task_type": "lexical_detection", "query": "try", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"try\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.6, 1.05, "try"]], "response_mode": "offline", "answer": "[{\"word\":\"try\",\"start\":0.6,\"end\":1.05}]"} {"utterance_id": "4507-16021-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0019.flac", "task_type": "lexical_detection", "query": "wretchedness", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"wretchedness\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.6, 2.42, "wretchedness"]], "response_mode": "offline", "answer": "[{\"word\":\"wretchedness\",\"start\":1.6,\"end\":2.42}]"} {"utterance_id": "7176-88083-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0013.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "After the audio ends, tally every instance of a body part. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[5.29, 5.63, "one"], [12.61, 12.7, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.29,\"end\":5.63},{\"word\":\"two\",\"start\":12.61,\"end\":12.7}]"} {"utterance_id": "4446-2275-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0021.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.93, 1.19, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":0.93,\"end\":1.19}]"} {"utterance_id": "1188-133604-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0013.flac", "task_type": "lexical_detection", "query": "remember", "prompt": "After the audio ends, extract all occurrences of the word \"remember\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.02, 1.41, "remember"]], "response_mode": "offline", "answer": "[{\"word\":\"remember\",\"start\":1.02,\"end\":1.41}]"} {"utterance_id": "7729-102255-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0032.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen to the full audio. After it finishes, find every instance of a food name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[18.28, 18.67, "dinner"]], "response_mode": "offline", "answer": "[{\"word\":\"dinner\",\"start\":18.28,\"end\":18.67}]"} {"utterance_id": "7021-85628-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0023.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[7.19, 7.71, "anders"]], "response_mode": "offline", "answer": "[{\"word\":\"anders\",\"start\":7.19,\"end\":7.71}]"} {"utterance_id": "1580-141083-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0045.flac", "task_type": "lexical_detection", "query": "suffering", "prompt": "When the utterance is complete, report all instances of the word \"suffering\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.82, 1.32, "suffering"]], "response_mode": "offline", "answer": "[{\"word\":\"suffering\",\"start\":0.82,\"end\":1.32}]"} {"utterance_id": "1221-135766-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0001.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen carefully to the entire clip, then report every instance of a family relation as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.5, 6.02, "child"], [9.81, 10.31, "parent"]], "response_mode": "offline", "answer": "[{\"word\":\"child\",\"start\":5.5,\"end\":6.02},{\"word\":\"parent\",\"start\":9.81,\"end\":10.31}]"} {"utterance_id": "6829-68769-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0001.flac", "task_type": "lexical_detection", "query": "state's", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"state's\" with fields word, start, end (seconds). If none, reply [].", "labels": [[6.41, 6.77, "state's"]], "response_mode": "offline", "answer": "[{\"word\":\"state's\",\"start\":6.41,\"end\":6.77}]"} {"utterance_id": "7127-75947-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0040.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for an animal name with word/start/end. If none, reply [].", "labels": [[7.79, 8.03, "one"], [9.21, 9.61, "two"], [10.9, 11.25, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.79,\"end\":8.03},{\"word\":\"two\",\"start\":9.21,\"end\":9.61},{\"word\":\"three\",\"start\":10.9,\"end\":11.25}]"} {"utterance_id": "1320-122617-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0015.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen to the full audio. After it finishes, find every instance of an animal name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.43, 0.91, "bear"]], "response_mode": "offline", "answer": "[{\"word\":\"bear\",\"start\":0.43,\"end\":0.91}]"} {"utterance_id": "121-127105-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0029.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.62, 5.17, "governess"]], "response_mode": "offline", "answer": "[{\"word\":\"governess\",\"start\":4.62,\"end\":5.17}]"} {"utterance_id": "4446-2275-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0016.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.43, 0.73, "hilda"]], "response_mode": "offline", "answer": "[{\"word\":\"hilda\",\"start\":0.43,\"end\":0.73}]"} {"utterance_id": "1089-134686-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0014.flac", "task_type": "lexical_detection", "query": "tried", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"tried\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.47, 0.76, "tried"]], "response_mode": "offline", "answer": "[{\"word\":\"tried\",\"start\":0.47,\"end\":0.76}]"} {"utterance_id": "1995-1826-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0002.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a plant name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.8, 4.32, "cotton"]], "response_mode": "offline", "answer": "[{\"word\":\"cotton\",\"start\":3.8,\"end\":4.32}]"} {"utterance_id": "7021-85628-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0027.flac", "task_type": "lexical_detection", "query": "dressed", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"dressed\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.67, 1.03, "dressed"]], "response_mode": "offline", "answer": "[{\"word\":\"dressed\",\"start\":0.67,\"end\":1.03}]"} {"utterance_id": "5142-33396-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0059.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.94, 4.22, "king"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":3.94,\"end\":4.22}]"} {"utterance_id": "672-122797-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0047.flac", "task_type": "lexical_detection", "query": "kind", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"kind\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.57, 1.26, "kind"]], "response_mode": "offline", "answer": "[{\"word\":\"kind\",\"start\":0.57,\"end\":1.26}]"} {"utterance_id": "121-123852-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123852/121-123852-0002.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.42, 2.82, "foot"]], "response_mode": "offline", "answer": "[{\"word\":\"foot\",\"start\":2.42,\"end\":2.82}]"} {"utterance_id": "2830-3979-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0007.flac", "task_type": "lexical_detection", "query": "instance", "prompt": "When the utterance is complete, report all instances of the word \"instance\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.72, 14.31, "instance"]], "response_mode": "offline", "answer": "[{\"word\":\"instance\",\"start\":13.72,\"end\":14.31}]"} {"utterance_id": "2094-142345-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0003.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a plant name with fields word, start, end (seconds). If none, reply [].", "labels": [[13.22, 13.54, "corn"]], "response_mode": "offline", "answer": "[{\"word\":\"corn\",\"start\":13.22,\"end\":13.54}]"} {"utterance_id": "7176-92135-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0040.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.32, 0.77, "hostess"]], "response_mode": "offline", "answer": "[{\"word\":\"hostess\",\"start\":0.32,\"end\":0.77}]"} {"utterance_id": "1580-141083-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0020.flac", "task_type": "lexical_detection", "query": "tiptoe", "prompt": "After the audio ends, extract all occurrences of the word \"tiptoe\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.1, 2.5, "tiptoe"]], "response_mode": "offline", "answer": "[{\"word\":\"tiptoe\",\"start\":2.1,\"end\":2.5}]"} {"utterance_id": "2094-142345-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0041.flac", "task_type": "lexical_detection", "query": "direction", "prompt": "When the utterance is complete, report all instances of the word \"direction\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 1.2, "direction"]], "response_mode": "offline", "answer": "[{\"word\":\"direction\",\"start\":0.51,\"end\":1.2}]"} {"utterance_id": "4507-16021-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0047.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Listen to the full audio. After it finishes, find every instance of a weather term. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[20.06, 20.56, "weather"]], "response_mode": "offline", "answer": "[{\"word\":\"weather\",\"start\":20.06,\"end\":20.56}]"} {"utterance_id": "4446-2275-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0012.flac", "task_type": "lexical_detection", "query": "forward", "prompt": "When the utterance is complete, report all instances of the word \"forward\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.79, 4.13, "forward"]], "response_mode": "offline", "answer": "[{\"word\":\"forward\",\"start\":3.79,\"end\":4.13}]"} {"utterance_id": "7127-75946-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0002.flac", "task_type": "lexical_counting", "query": "king", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"king\" (word, start, end). If none, reply [].", "labels": [[1.94, 2.41, "one"], [10.24, 10.48, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.94,\"end\":2.41},{\"word\":\"two\",\"start\":10.24,\"end\":10.48}]"} {"utterance_id": "6829-68771-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0009.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, list every instance of an organization name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.23, 4.16, "organization"]], "response_mode": "offline", "answer": "[{\"word\":\"organization\",\"start\":3.23,\"end\":4.16}]"} {"utterance_id": "4970-29095-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0011.flac", "task_type": "lexical_detection", "query": "stand", "prompt": "After listening to the whole recording, output every instance of the word \"stand\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.74, 2.16, "stand"]], "response_mode": "offline", "answer": "[{\"word\":\"stand\",\"start\":1.74,\"end\":2.16}]"} {"utterance_id": "260-123286-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0027.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Once the audio has finished, list each occurrence of a body part as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[1.2, 1.54, "one"], [6.46, 6.96, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.2,\"end\":1.54},{\"word\":\"two\",\"start\":6.46,\"end\":6.96}]"} {"utterance_id": "8463-294825-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0002.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.39, 0.82, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":0.39,\"end\":0.82}]"} {"utterance_id": "6930-76324-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0015.flac", "task_type": "lexical_detection", "query": "dust", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"dust\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.48, 10.83, "dust"]], "response_mode": "offline", "answer": "[{\"word\":\"dust\",\"start\":10.48,\"end\":10.83}]"} {"utterance_id": "6829-68771-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0006.flac", "task_type": "lexical_detection", "query": "beth", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"beth\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.9, 2.33, "beth"]], "response_mode": "offline", "answer": "[{\"word\":\"beth\",\"start\":1.9,\"end\":2.33}]"} {"utterance_id": "7176-92135-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0025.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.85, 1.3, "hamlet"]], "response_mode": "offline", "answer": "[{\"word\":\"hamlet\",\"start\":0.85,\"end\":1.3}]"} {"utterance_id": "6930-81414-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0026.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "When the utterance is complete, report all instances of an ordinal number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.42, 2.95, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":2.42,\"end\":2.95}]"} {"utterance_id": "61-70970-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.15, 1.52, "robin"]], "response_mode": "offline", "answer": "[{\"word\":\"robin\",\"start\":1.15,\"end\":1.52}]"} {"utterance_id": "1320-122617-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0035.flac", "task_type": "lexical_detection", "query": "heaving", "prompt": "When the utterance is complete, report all instances of the word \"heaving\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.8, 1.18, "heaving"]], "response_mode": "offline", "answer": "[{\"word\":\"heaving\",\"start\":0.8,\"end\":1.18}]"} {"utterance_id": "8463-287645-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0008.flac", "task_type": "lexical_detection", "query": "usual", "prompt": "After listening to the whole recording, output every instance of the word \"usual\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.56, 0.98, "usual"]], "response_mode": "offline", "answer": "[{\"word\":\"usual\",\"start\":0.56,\"end\":0.98}]"} {"utterance_id": "8230-279154-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0039.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.05, 2.26, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.05,\"end\":2.26}]"} {"utterance_id": "5105-28241-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0017.flac", "task_type": "lexical_counting", "query": "fathoms", "prompt": "Process the full audio offline. For every occurrence of the word \"fathoms\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.21, 8.84, "one"], [9.96, 10.7, "two"], [11.73, 12.48, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":8.21,\"end\":8.84},{\"word\":\"two\",\"start\":9.96,\"end\":10.7},{\"word\":\"three\",\"start\":11.73,\"end\":12.48}]"} {"utterance_id": "1580-141084-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0033.flac", "task_type": "lexical_detection", "query": "criminal", "prompt": "After listening to the whole recording, output every instance of the word \"criminal\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.23, 6.66, "criminal"]], "response_mode": "offline", "answer": "[{\"word\":\"criminal\",\"start\":6.23,\"end\":6.66}]"} {"utterance_id": "237-134500-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0004.flac", "task_type": "lexical_detection", "query": "invitation", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"invitation\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.72, 1.33, "invitation"]], "response_mode": "offline", "answer": "[{\"word\":\"invitation\",\"start\":0.72,\"end\":1.33}]"} {"utterance_id": "7729-102255-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0042.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.76, 3.46, "robinson"]], "response_mode": "offline", "answer": "[{\"word\":\"robinson\",\"start\":2.76,\"end\":3.46}]"} {"utterance_id": "7127-75946-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0011.flac", "task_type": "lexical_detection", "query": "private", "prompt": "After the audio ends, extract all occurrences of the word \"private\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.97, 2.41, "private"]], "response_mode": "offline", "answer": "[{\"word\":\"private\",\"start\":1.97,\"end\":2.41}]"} {"utterance_id": "2830-3980-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0036.flac", "task_type": "lexical_detection", "query": "holy", "prompt": "After listening to the whole recording, output every instance of the word \"holy\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.55, 2.85, "holy"]], "response_mode": "offline", "answer": "[{\"word\":\"holy\",\"start\":2.55,\"end\":2.85}]"} {"utterance_id": "4446-2273-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.91, 2.39, "hilda"]], "response_mode": "offline", "answer": "[{\"word\":\"hilda\",\"start\":1.91,\"end\":2.39}]"} {"utterance_id": "672-122797-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0013.flac", "task_type": "lexical_detection", "query": "spread", "prompt": "When the utterance is complete, report all instances of the word \"spread\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.4, 4.07, "spread"]], "response_mode": "offline", "answer": "[{\"word\":\"spread\",\"start\":3.4,\"end\":4.07}]"} {"utterance_id": "5105-28241-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0004.flac", "task_type": "lexical_detection", "query": "moderate", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"moderate\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.14, 1.68, "moderate"]], "response_mode": "offline", "answer": "[{\"word\":\"moderate\",\"start\":1.14,\"end\":1.68}]"} {"utterance_id": "3570-5694-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0004.flac", "task_type": "lexical_detection", "query": "luxuries", "prompt": "After listening to the whole recording, output every instance of the word \"luxuries\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.76, 2.43, "luxuries"]], "response_mode": "offline", "answer": "[{\"word\":\"luxuries\",\"start\":1.76,\"end\":2.43}]"} {"utterance_id": "6930-76324-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0000.flac", "task_type": "lexical_detection", "query": "goliath", "prompt": "Once the audio has finished, identify every instance of the word \"goliath\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.37, 1.0, "goliath"]], "response_mode": "offline", "answer": "[{\"word\":\"goliath\",\"start\":0.37,\"end\":1.0}]"} {"utterance_id": "1320-122617-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0014.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.61, 4.11, "conjurer"]], "response_mode": "offline", "answer": "[{\"word\":\"conjurer\",\"start\":3.61,\"end\":4.11}]"} {"utterance_id": "5683-32865-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0006.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a food name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.74, 1.2, "dinner"]], "response_mode": "offline", "answer": "[{\"word\":\"dinner\",\"start\":0.74,\"end\":1.2}]"} {"utterance_id": "3729-6852-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0029.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen carefully to the entire clip, then report every instance of an animal name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.08, 8.08, "civet cat"]], "response_mode": "offline", "answer": "[{\"word\":\"civet cat\",\"start\":7.08,\"end\":8.08}]"} {"utterance_id": "8555-284447-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0002.flac", "task_type": "lexical_detection", "query": "mornin", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"mornin\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.41, 5.93, "mornin"]], "response_mode": "offline", "answer": "[{\"word\":\"mornin\",\"start\":5.41,\"end\":5.93}]"} {"utterance_id": "4446-2275-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0031.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.24, 1.71, "bartley"]], "response_mode": "offline", "answer": "[{\"word\":\"bartley\",\"start\":1.24,\"end\":1.71}]"} {"utterance_id": "4446-2275-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 1.05, "alexander"]], "response_mode": "offline", "answer": "[{\"word\":\"alexander\",\"start\":0.52,\"end\":1.05}]"} {"utterance_id": "1995-1826-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0008.flac", "task_type": "lexical_detection", "query": "county", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"county\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.0, 2.47, "county"]], "response_mode": "offline", "answer": "[{\"word\":\"county\",\"start\":2.0,\"end\":2.47}]"} {"utterance_id": "260-123286-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0018.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Listen to the full audio. Count every occurrence of a number. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.02, 3.17, "one"], [4.29, 4.55, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.02,\"end\":3.17},{\"word\":\"two\",\"start\":4.29,\"end\":4.55}]"} {"utterance_id": "5142-33396-0067", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0067.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.0, 1.36, "tongue"]], "response_mode": "offline", "answer": "[{\"word\":\"tongue\",\"start\":1.0,\"end\":1.36}]"} {"utterance_id": "237-126133-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0016.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.84, 1.33, "jasper"]], "response_mode": "offline", "answer": "[{\"word\":\"jasper\",\"start\":0.84,\"end\":1.33}]"} {"utterance_id": "5142-33396-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0002.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.25, 0.87, "two hundred"]], "response_mode": "offline", "answer": "[{\"word\":\"two hundred\",\"start\":0.25,\"end\":0.87}]"} {"utterance_id": "4970-29095-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0007.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the full audio. After it finishes, find every instance of an ordinal number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.13, 8.57, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":8.13,\"end\":8.57}]"} {"utterance_id": "1188-133604-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0038.flac", "task_type": "lexical_detection", "query": "turner", "prompt": "Once the audio has finished, identify every instance of the word \"turner\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.85, 4.2, "turner"]], "response_mode": "offline", "answer": "[{\"word\":\"turner\",\"start\":3.85,\"end\":4.2}]"} {"utterance_id": "4077-13751-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0008.flac", "task_type": "lexical_detection", "query": "regarded", "prompt": "After the audio ends, extract all occurrences of the word \"regarded\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.62, 3.09, "regarded"]], "response_mode": "offline", "answer": "[{\"word\":\"regarded\",\"start\":2.62,\"end\":3.09}]"} {"utterance_id": "2300-131720-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0004.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "After the audio ends, extract all occurrences of a state or province name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.09, 6.72, "districts"]], "response_mode": "offline", "answer": "[{\"word\":\"districts\",\"start\":6.09,\"end\":6.72}]"} {"utterance_id": "1284-1180-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0017.flac", "task_type": "semantic_counting", "query": "color", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a color name (word, start, end). If none, reply [].", "labels": [[2.67, 2.92, "one"], [4.25, 4.55, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.67,\"end\":2.92},{\"word\":\"two\",\"start\":4.25,\"end\":4.55}]"} {"utterance_id": "4992-41797-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0004.flac", "task_type": "lexical_detection", "query": "doctor", "prompt": "Once the audio has finished, identify every instance of the word \"doctor\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.91, 6.33, "doctor"]], "response_mode": "offline", "answer": "[{\"word\":\"doctor\",\"start\":5.91,\"end\":6.33}]"} {"utterance_id": "4970-29093-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0006.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen carefully to the entire clip, then report every instance of a money or currency term as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[28.84, 29.22, "fees"]], "response_mode": "offline", "answer": "[{\"word\":\"fees\",\"start\":28.84,\"end\":29.22}]"} {"utterance_id": "2830-3980-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0006.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.19, 3.43, "paul"]], "response_mode": "offline", "answer": "[{\"word\":\"paul\",\"start\":3.19,\"end\":3.43}]"} {"utterance_id": "8463-294825-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0009.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Do not answer until the audio ends. Then list each instance of an animal name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[11.98, 12.54, "animal"]], "response_mode": "offline", "answer": "[{\"word\":\"animal\",\"start\":11.98,\"end\":12.54}]"} {"utterance_id": "8224-274381-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0002.flac", "task_type": "lexical_detection", "query": "foretold", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"foretold\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.39, 2.35, "foretold"]], "response_mode": "offline", "answer": "[{\"word\":\"foretold\",\"start\":1.39,\"end\":2.35}]"} {"utterance_id": "4446-2273-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0008.flac", "task_type": "lexical_detection", "query": "mike", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"mike\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.22, 6.54, "mike"]], "response_mode": "offline", "answer": "[{\"word\":\"mike\",\"start\":6.22,\"end\":6.54}]"} {"utterance_id": "7021-85628-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0023.flac", "task_type": "lexical_detection", "query": "king's", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"king's\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.25, 4.64, "king's"]], "response_mode": "offline", "answer": "[{\"word\":\"king's\",\"start\":4.25,\"end\":4.64}]"} {"utterance_id": "8230-279154-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0038.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an ordinal number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.65, 12.13, "first"], [20.28, 20.76, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":11.65,\"end\":12.13},{\"word\":\"first\",\"start\":20.28,\"end\":20.76}]"} {"utterance_id": "7176-88083-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0003.flac", "task_type": "lexical_detection", "query": "claws", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"claws\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.06, 7.46, "claws"]], "response_mode": "offline", "answer": "[{\"word\":\"claws\",\"start\":7.06,\"end\":7.46}]"} {"utterance_id": "672-122797-0061", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0061.flac", "task_type": "lexical_detection", "query": "candles", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"candles\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.06, 4.03, "candles"]], "response_mode": "offline", "answer": "[{\"word\":\"candles\",\"start\":3.06,\"end\":4.03}]"} {"utterance_id": "7176-92135-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0014.flac", "task_type": "lexical_detection", "query": "granted", "prompt": "When the utterance is complete, report all instances of the word \"granted\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.61, 1.06, "granted"]], "response_mode": "offline", "answer": "[{\"word\":\"granted\",\"start\":0.61,\"end\":1.06}]"} {"utterance_id": "672-122797-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0054.flac", "task_type": "lexical_detection", "query": "tree", "prompt": "Once the audio has finished, identify every instance of the word \"tree\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.38, 3.86, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":3.38,\"end\":3.86}]"} {"utterance_id": "908-157963-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0025.flac", "task_type": "lexical_counting", "query": "none", "prompt": "Listen to the full audio. Count every occurrence of the word \"none\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.39, 4.64, "one"], [5.66, 5.97, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.39,\"end\":4.64},{\"word\":\"two\",\"start\":5.66,\"end\":5.97}]"} {"utterance_id": "6829-68771-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0002.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Process the full audio offline. Find every instance of an organization name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.64, 8.52, "opposition"]], "response_mode": "offline", "answer": "[{\"word\":\"opposition\",\"start\":7.64,\"end\":8.52}]"} {"utterance_id": "2094-142345-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0011.flac", "task_type": "lexical_detection", "query": "complexion", "prompt": "When the utterance is complete, report all instances of the word \"complexion\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.61, 12.36, "complexion"]], "response_mode": "offline", "answer": "[{\"word\":\"complexion\",\"start\":11.61,\"end\":12.36}]"} {"utterance_id": "121-121726-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0012.flac", "task_type": "lexical_detection", "query": "bond", "prompt": "After listening to the whole recording, output every instance of the word \"bond\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.06, 2.63, "bond"]], "response_mode": "offline", "answer": "[{\"word\":\"bond\",\"start\":2.06,\"end\":2.63}]"} {"utterance_id": "6829-68769-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0029.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, list every instance of an organization name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.87, 1.35, "company"]], "response_mode": "offline", "answer": "[{\"word\":\"company\",\"start\":0.87,\"end\":1.35}]"} {"utterance_id": "6829-68769-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0050.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.32, 0.68, "beth"]], "response_mode": "offline", "answer": "[{\"word\":\"beth\",\"start\":0.32,\"end\":0.68}]"} {"utterance_id": "4970-29095-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0004.flac", "task_type": "lexical_detection", "query": "whipped", "prompt": "Process the full audio offline. Find every instance of the word \"whipped\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.78, 3.07, "whipped"]], "response_mode": "offline", "answer": "[{\"word\":\"whipped\",\"start\":2.78,\"end\":3.07}]"} {"utterance_id": "4970-29093-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0003.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Process the full audio offline. Find every instance of an organization name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.29, 10.04, "newspaper"]], "response_mode": "offline", "answer": "[{\"word\":\"newspaper\",\"start\":9.29,\"end\":10.04}]"} {"utterance_id": "260-123288-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0011.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.39, 5.71, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":5.39,\"end\":5.71}]"} {"utterance_id": "237-134500-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0007.flac", "task_type": "lexical_detection", "query": "never", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"never\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.22, 0.47, "never"]], "response_mode": "offline", "answer": "[{\"word\":\"never\",\"start\":0.22,\"end\":0.47}]"} {"utterance_id": "1188-133604-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0032.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "After listening to the whole recording, output timed count words for a person's name (one, two, three, ...) as JSON. If none, reply [].", "labels": [[2.46, 3.7, "one"], [5.38, 6.01, "two"], [8.11, 8.68, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.46,\"end\":3.7},{\"word\":\"two\",\"start\":5.38,\"end\":6.01},{\"word\":\"three\",\"start\":8.11,\"end\":8.68}]"} {"utterance_id": "5142-36377-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0021.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.79, 11.38, "naomi"], [15.92, 16.5, "ambrose"]], "response_mode": "offline", "answer": "[{\"word\":\"naomi\",\"start\":10.79,\"end\":11.38},{\"word\":\"ambrose\",\"start\":15.92,\"end\":16.5}]"} {"utterance_id": "4970-29095-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0018.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.53, 0.84, "ruth"]], "response_mode": "offline", "answer": "[{\"word\":\"ruth\",\"start\":0.53,\"end\":0.84}]"} {"utterance_id": "4970-29093-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0012.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.77, 1.1, "philip"]], "response_mode": "offline", "answer": "[{\"word\":\"philip\",\"start\":0.77,\"end\":1.1}]"} {"utterance_id": "1580-141084-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0045.flac", "task_type": "lexical_detection", "query": "escape", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"escape\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.96, 3.44, "escape"]], "response_mode": "offline", "answer": "[{\"word\":\"escape\",\"start\":2.96,\"end\":3.44}]"} {"utterance_id": "1089-134686-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0021.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a food name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.68, 3.1, "water"]], "response_mode": "offline", "answer": "[{\"word\":\"water\",\"start\":2.68,\"end\":3.1}]"} {"utterance_id": "4507-16021-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0032.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.97, 2.36, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":1.97,\"end\":2.36}]"} {"utterance_id": "7021-79730-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0001.flac", "task_type": "lexical_detection", "query": "aid", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"aid\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.22, 4.67, "aid"]], "response_mode": "offline", "answer": "[{\"word\":\"aid\",\"start\":4.22,\"end\":4.67}]"} {"utterance_id": "4446-2273-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0010.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "You will hear a complete utterance. When it ends, list every instance of a plant name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.63, 1.27, "watercress"], [7.54, 8.3, "artichokes"]], "response_mode": "offline", "answer": "[{\"word\":\"watercress\",\"start\":0.63,\"end\":1.27},{\"word\":\"artichokes\",\"start\":7.54,\"end\":8.3}]"} {"utterance_id": "4446-2275-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.35, 0.86, "alexander"], [6.44, 6.81, "hilda's"]], "response_mode": "offline", "answer": "[{\"word\":\"alexander\",\"start\":0.35,\"end\":0.86},{\"word\":\"hilda's\",\"start\":6.44,\"end\":6.81}]"} {"utterance_id": "7729-102255-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0020.flac", "task_type": "lexical_detection", "query": "posse", "prompt": "Once the audio has finished, identify every instance of the word \"posse\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.58, 4.05, "posse"]], "response_mode": "offline", "answer": "[{\"word\":\"posse\",\"start\":3.58,\"end\":4.05}]"} {"utterance_id": "3575-170457-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0026.flac", "task_type": "lexical_counting", "query": "writing", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"writing\" with word/start/end. If none, reply [].", "labels": [[2.96, 3.3, "one"], [5.35, 5.74, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.96,\"end\":3.3},{\"word\":\"two\",\"start\":5.35,\"end\":5.74}]"} {"utterance_id": "4077-13751-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0011.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an ordinal number with fields word, start, end (seconds). If none, reply [].", "labels": [[8.49, 9.08, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":8.49,\"end\":9.08}]"} {"utterance_id": "260-123440-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0010.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Once the audio has finished, identify every instance of an animal name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.27, 5.84, "fishes"]], "response_mode": "offline", "answer": "[{\"word\":\"fishes\",\"start\":5.27,\"end\":5.84}]"} {"utterance_id": "1089-134686-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0036.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.46, 3.08, "saint francis xavier"]], "response_mode": "offline", "answer": "[{\"word\":\"saint francis xavier\",\"start\":1.46,\"end\":3.08}]"} {"utterance_id": "2830-3980-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0022.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Do not answer until the audio ends. Then list each instance of an ordinal number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.41, 2.75, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":2.41,\"end\":2.75}]"} {"utterance_id": "8224-274384-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0006.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.63, 1.0, "king"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":0.63,\"end\":1.0}]"} {"utterance_id": "4970-29093-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0014.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.76, 2.4, "engineer"]], "response_mode": "offline", "answer": "[{\"word\":\"engineer\",\"start\":1.76,\"end\":2.4}]"} {"utterance_id": "1221-135766-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0000.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.73, 11.3, "features"]], "response_mode": "offline", "answer": "[{\"word\":\"features\",\"start\":10.73,\"end\":11.3}]"} {"utterance_id": "7729-102255-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0016.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Once the audio has finished, identify every instance of a state or province name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.81, 5.24, "missouri"], [6.32, 6.84, "kansas"]], "response_mode": "offline", "answer": "[{\"word\":\"missouri\",\"start\":4.81,\"end\":5.24},{\"word\":\"kansas\",\"start\":6.32,\"end\":6.84}]"} {"utterance_id": "5683-32865-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0009.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.88, 2.41, "fingers"]], "response_mode": "offline", "answer": "[{\"word\":\"fingers\",\"start\":1.88,\"end\":2.41}]"} {"utterance_id": "2830-3980-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0003.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.23, 0.54, "paul"]], "response_mode": "offline", "answer": "[{\"word\":\"paul\",\"start\":0.23,\"end\":0.54}]"} {"utterance_id": "1995-1826-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0014.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the full audio. After it finishes, find every instance of a plant name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.22, 1.0, "cotton"]], "response_mode": "offline", "answer": "[{\"word\":\"cotton\",\"start\":0.22,\"end\":1.0}]"} {"utterance_id": "8455-210777-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0040.flac", "task_type": "lexical_detection", "query": "ferdinando", "prompt": "After listening to the whole recording, output every instance of the word \"ferdinando\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.84, 2.67, "ferdinando"]], "response_mode": "offline", "answer": "[{\"word\":\"ferdinando\",\"start\":1.84,\"end\":2.67}]"} {"utterance_id": "1188-133604-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0019.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[9.38, 9.88, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":9.38,\"end\":9.88}]"} {"utterance_id": "1995-1826-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0015.flac", "task_type": "lexical_detection", "query": "forgotten", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"forgotten\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.87, 1.4, "forgotten"]], "response_mode": "offline", "answer": "[{\"word\":\"forgotten\",\"start\":0.87,\"end\":1.4}]"} {"utterance_id": "1089-134686-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0005.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.18, 4.74, "shelley's"]], "response_mode": "offline", "answer": "[{\"word\":\"shelley's\",\"start\":4.18,\"end\":4.74}]"} {"utterance_id": "260-123288-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0009.flac", "task_type": "lexical_detection", "query": "crush", "prompt": "Process the full audio offline. Find every instance of the word \"crush\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.76, 2.18, "crush"]], "response_mode": "offline", "answer": "[{\"word\":\"crush\",\"start\":1.76,\"end\":2.18}]"} {"utterance_id": "4077-13754-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0012.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[6.38, 6.71, "six"], [8.65, 9.22, "three hundred"], [20.53, 21.02, "millions"]], "response_mode": "offline", "answer": "[{\"word\":\"six\",\"start\":6.38,\"end\":6.71},{\"word\":\"three hundred\",\"start\":8.65,\"end\":9.22},{\"word\":\"millions\",\"start\":20.53,\"end\":21.02}]"} {"utterance_id": "8463-294828-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0006.flac", "task_type": "lexical_detection", "query": "gardens", "prompt": "When the utterance is complete, report all instances of the word \"gardens\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.23, 4.75, "gardens"]], "response_mode": "offline", "answer": "[{\"word\":\"gardens\",\"start\":4.23,\"end\":4.75}]"} {"utterance_id": "237-134500-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0024.flac", "task_type": "semantic_detection", "query": "city", "prompt": "After the audio ends, extract all occurrences of a city name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.95, 2.52, "new york"]], "response_mode": "offline", "answer": "[{\"word\":\"new york\",\"start\":1.95,\"end\":2.52}]"} {"utterance_id": "6930-75918-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0001.flac", "task_type": "lexical_counting", "query": "french", "prompt": "Process the full audio offline. For every occurrence of the word \"french\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.67, 2.08, "one"], [8.52, 9.01, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.67,\"end\":2.08},{\"word\":\"two\",\"start\":8.52,\"end\":9.01}]"} {"utterance_id": "6930-76324-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.41, 0.79, "joyce"]], "response_mode": "offline", "answer": "[{\"word\":\"joyce\",\"start\":0.41,\"end\":0.79}]"} {"utterance_id": "7021-85628-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0019.flac", "task_type": "lexical_detection", "query": "jump", "prompt": "After listening to the whole recording, output every instance of the word \"jump\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.82, 1.23, "jump"]], "response_mode": "offline", "answer": "[{\"word\":\"jump\",\"start\":0.82,\"end\":1.23}]"} {"utterance_id": "6930-75918-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0018.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "After listening to the whole recording, output every instance of a time of day with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.64, 7.04, "evening"]], "response_mode": "offline", "answer": "[{\"word\":\"evening\",\"start\":6.64,\"end\":7.04}]"} {"utterance_id": "1995-1837-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0022.flac", "task_type": "lexical_detection", "query": "sick", "prompt": "After listening to the whole recording, output every instance of the word \"sick\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.65, 0.97, "sick"]], "response_mode": "offline", "answer": "[{\"word\":\"sick\",\"start\":0.65,\"end\":0.97}]"} {"utterance_id": "4992-41797-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0006.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Do not answer until the audio ends. Then list each instance of an animal name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.36, 3.64, "hen"]], "response_mode": "offline", "answer": "[{\"word\":\"hen\",\"start\":3.36,\"end\":3.64}]"} {"utterance_id": "7127-75946-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0013.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.6, 0.89, "king"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":0.6,\"end\":0.89}]"} {"utterance_id": "3575-170457-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0046.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a person's name (word, start, end). If none, reply [].", "labels": [[1.36, 2.15, "one"], [24.03, 24.47, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.36,\"end\":2.15},{\"word\":\"two\",\"start\":24.03,\"end\":24.47}]"} {"utterance_id": "4446-2275-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0041.flac", "task_type": "lexical_detection", "query": "makes", "prompt": "After listening to the whole recording, output every instance of the word \"makes\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.68, 3.0, "makes"]], "response_mode": "offline", "answer": "[{\"word\":\"makes\",\"start\":2.68,\"end\":3.0}]"} {"utterance_id": "8455-210777-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0038.flac", "task_type": "lexical_detection", "query": "receive", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"receive\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.88, 6.34, "receive"]], "response_mode": "offline", "answer": "[{\"word\":\"receive\",\"start\":5.88,\"end\":6.34}]"} {"utterance_id": "4077-13754-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0002.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Do not answer until the audio ends. Then list each instance of a state or province name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.06, 4.59, "utah"]], "response_mode": "offline", "answer": "[{\"word\":\"utah\",\"start\":4.06,\"end\":4.59}]"} {"utterance_id": "6829-68771-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0010.flac", "task_type": "semantic_detection", "query": "city", "prompt": "When the utterance is complete, report all instances of a city name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.5, 0.97, "fairview"]], "response_mode": "offline", "answer": "[{\"word\":\"fairview\",\"start\":0.5,\"end\":0.97}]"} {"utterance_id": "7176-88083-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0023.flac", "task_type": "lexical_detection", "query": "drop", "prompt": "When the utterance is complete, report all instances of the word \"drop\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.93, 6.26, "drop"]], "response_mode": "offline", "answer": "[{\"word\":\"drop\",\"start\":5.93,\"end\":6.26}]"} {"utterance_id": "8455-210777-0061", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0061.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.3, 10.19, "mister neverbend"]], "response_mode": "offline", "answer": "[{\"word\":\"mister neverbend\",\"start\":9.3,\"end\":10.19}]"} {"utterance_id": "8230-279154-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0008.flac", "task_type": "lexical_detection", "query": "warranted", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"warranted\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.6, 3.13, "warranted"]], "response_mode": "offline", "answer": "[{\"word\":\"warranted\",\"start\":2.6,\"end\":3.13}]"} {"utterance_id": "1221-135767-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0019.flac", "task_type": "lexical_detection", "query": "cried", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"cried\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.58, 0.9, "cried"]], "response_mode": "offline", "answer": "[{\"word\":\"cried\",\"start\":0.58,\"end\":0.9}]"} {"utterance_id": "8224-274381-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0015.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[11.4, 12.3, "commanders"]], "response_mode": "offline", "answer": "[{\"word\":\"commanders\",\"start\":11.4,\"end\":12.3}]"} {"utterance_id": "2830-3980-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0009.flac", "task_type": "lexical_detection", "query": "cetera", "prompt": "Once the audio has finished, identify every instance of the word \"cetera\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.95, 3.42, "cetera"]], "response_mode": "offline", "answer": "[{\"word\":\"cetera\",\"start\":2.95,\"end\":3.42}]"} {"utterance_id": "7127-75946-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0007.flac", "task_type": "lexical_detection", "query": "comply", "prompt": "Process the full audio offline. Find every instance of the word \"comply\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.24, 2.82, "comply"]], "response_mode": "offline", "answer": "[{\"word\":\"comply\",\"start\":2.24,\"end\":2.82}]"} {"utterance_id": "5683-32865-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0005.flac", "task_type": "lexical_detection", "query": "frozen", "prompt": "After listening to the whole recording, output every instance of the word \"frozen\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.37, 5.05, "frozen"]], "response_mode": "offline", "answer": "[{\"word\":\"frozen\",\"start\":4.37,\"end\":5.05}]"} {"utterance_id": "7127-75946-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0017.flac", "task_type": "lexical_counting", "query": "flesh", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"flesh\" (word, start, end). If none, reply [].", "labels": [[5.36, 5.68, "one"], [11.14, 11.45, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.36,\"end\":5.68},{\"word\":\"two\",\"start\":11.14,\"end\":11.45}]"} {"utterance_id": "4507-16021-0057", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0057.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "After listening to the whole recording, output every instance of a disease name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.3, 4.17, "burns"]], "response_mode": "offline", "answer": "[{\"word\":\"burns\",\"start\":3.3,\"end\":4.17}]"} {"utterance_id": "3729-6852-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0038.flac", "task_type": "lexical_detection", "query": "fortunate", "prompt": "After the audio ends, extract all occurrences of the word \"fortunate\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.6, 4.38, "fortunate"]], "response_mode": "offline", "answer": "[{\"word\":\"fortunate\",\"start\":3.6,\"end\":4.38}]"} {"utterance_id": "8463-294828-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0037.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.02, 2.52, "throats"]], "response_mode": "offline", "answer": "[{\"word\":\"throats\",\"start\":2.02,\"end\":2.52}]"} {"utterance_id": "2094-142345-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0008.flac", "task_type": "lexical_detection", "query": "sparkles", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"sparkles\" with fields word, start, end (seconds). If none, reply [].", "labels": [[9.51, 10.25, "sparkles"]], "response_mode": "offline", "answer": "[{\"word\":\"sparkles\",\"start\":9.51,\"end\":10.25}]"} {"utterance_id": "6829-68769-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0015.flac", "task_type": "lexical_detection", "query": "yearning", "prompt": "When the utterance is complete, report all instances of the word \"yearning\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.78, 2.13, "yearning"]], "response_mode": "offline", "answer": "[{\"word\":\"yearning\",\"start\":1.78,\"end\":2.13}]"} {"utterance_id": "4446-2271-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0005.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Do not answer until the audio ends. Then list each instance of an ordinal number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.69, 3.06, "second"]], "response_mode": "offline", "answer": "[{\"word\":\"second\",\"start\":2.69,\"end\":3.06}]"} {"utterance_id": "1284-1180-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0012.flac", "task_type": "lexical_counting", "query": "place", "prompt": "After listening to the whole recording, output timed count words for the word \"place\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[1.77, 2.1, "one"], [4.21, 4.8, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.77,\"end\":2.1},{\"word\":\"two\",\"start\":4.21,\"end\":4.8}]"} {"utterance_id": "5683-32879-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0017.flac", "task_type": "lexical_detection", "query": "promise", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"promise\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.96, 9.48, "promise"]], "response_mode": "offline", "answer": "[{\"word\":\"promise\",\"start\":8.96,\"end\":9.48}]"} {"utterance_id": "1284-1180-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0022.flac", "task_type": "lexical_detection", "query": "afraid", "prompt": "After listening to the whole recording, output every instance of the word \"afraid\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 0.81, "afraid"]], "response_mode": "offline", "answer": "[{\"word\":\"afraid\",\"start\":0.45,\"end\":0.81}]"} {"utterance_id": "5142-33396-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0018.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.26, 0.78, "dragon's"]], "response_mode": "offline", "answer": "[{\"word\":\"dragon's\",\"start\":0.26,\"end\":0.78}]"} {"utterance_id": "5105-28240-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0011.flac", "task_type": "semantic_detection", "query": "country", "prompt": "When the utterance is complete, report all instances of a country name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.28, 1.97, "continent"]], "response_mode": "offline", "answer": "[{\"word\":\"continent\",\"start\":1.28,\"end\":1.97}]"} {"utterance_id": "4507-16021-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0027.flac", "task_type": "lexical_detection", "query": "good", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"good\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.35, 1.89, "good"]], "response_mode": "offline", "answer": "[{\"word\":\"good\",\"start\":1.35,\"end\":1.89}]"} {"utterance_id": "4446-2273-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0008.flac", "task_type": "semantic_counting", "query": "kinship", "prompt": "After listening to the whole recording, output timed count words for a family relation (one, two, three, ...) as JSON. If none, reply [].", "labels": [[3.63, 4.11, "one"], [5.9, 6.22, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.63,\"end\":4.11},{\"word\":\"two\",\"start\":5.9,\"end\":6.22}]"} {"utterance_id": "1320-122617-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0037.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Once the audio has finished, identify every instance of an animal name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.04, 1.39, "dog"]], "response_mode": "offline", "answer": "[{\"word\":\"dog\",\"start\":1.04,\"end\":1.39}]"} {"utterance_id": "121-121726-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0014.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "When the utterance is complete, report all instances of an animal name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.96, 2.44, "horse"]], "response_mode": "offline", "answer": "[{\"word\":\"horse\",\"start\":1.96,\"end\":2.44}]"} {"utterance_id": "8224-274384-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0011.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.89, 6.39, "king"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":5.89,\"end\":6.39}]"} {"utterance_id": "8555-284447-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0023.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.71, 4.99, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.71,\"end\":4.99}]"} {"utterance_id": "237-134500-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0013.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Once the audio has finished, identify every instance of a plant name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.1, 1.69, "cherries"], [4.94, 5.37, "branch"]], "response_mode": "offline", "answer": "[{\"word\":\"cherries\",\"start\":1.1,\"end\":1.69},{\"word\":\"branch\",\"start\":4.94,\"end\":5.37}]"} {"utterance_id": "4077-13751-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0016.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.14, 2.74, "officers"]], "response_mode": "offline", "answer": "[{\"word\":\"officers\",\"start\":2.14,\"end\":2.74}]"} {"utterance_id": "237-126133-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.34, 2.05, "missus whitney"]], "response_mode": "offline", "answer": "[{\"word\":\"missus whitney\",\"start\":1.34,\"end\":2.05}]"} {"utterance_id": "237-134500-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0016.flac", "task_type": "lexical_detection", "query": "know", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"know\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.79, 1.01, "know"], [2.04, 2.23, "know"]], "response_mode": "offline", "answer": "[{\"word\":\"know\",\"start\":0.79,\"end\":1.01},{\"word\":\"know\",\"start\":2.04,\"end\":2.23}]"} {"utterance_id": "5683-32866-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0000.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.31, 0.9, "miss lake"]], "response_mode": "offline", "answer": "[{\"word\":\"miss lake\",\"start\":0.31,\"end\":0.9}]"} {"utterance_id": "1580-141084-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0022.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Process the full audio offline. Find every instance of a time of day and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.33, 1.72, "morning"]], "response_mode": "offline", "answer": "[{\"word\":\"morning\",\"start\":1.33,\"end\":1.72}]"} {"utterance_id": "1221-135766-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0010.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.08, 14.64, "child"]], "response_mode": "offline", "answer": "[{\"word\":\"child\",\"start\":14.08,\"end\":14.64}]"} {"utterance_id": "1320-122617-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0017.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.68, 3.01, "scout"]], "response_mode": "offline", "answer": "[{\"word\":\"scout\",\"start\":2.68,\"end\":3.01}]"} {"utterance_id": "4970-29093-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0005.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[11.76, 12.33, "missionary"]], "response_mode": "offline", "answer": "[{\"word\":\"missionary\",\"start\":11.76,\"end\":12.33}]"} {"utterance_id": "4446-2273-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.45, 1.76, "molly"]], "response_mode": "offline", "answer": "[{\"word\":\"molly\",\"start\":1.45,\"end\":1.76}]"} {"utterance_id": "1188-133604-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0044.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Listen to the full audio. After it finishes, find every instance of a weather term. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.05, 8.42, "cloud"]], "response_mode": "offline", "answer": "[{\"word\":\"cloud\",\"start\":8.05,\"end\":8.42}]"} {"utterance_id": "2094-142345-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0051.flac", "task_type": "lexical_detection", "query": "message", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"message\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.32, 3.68, "message"]], "response_mode": "offline", "answer": "[{\"word\":\"message\",\"start\":3.32,\"end\":3.68}]"} {"utterance_id": "121-123852-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123852/121-123852-0002.flac", "task_type": "lexical_detection", "query": "farthest", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"farthest\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.52, 5.2, "farthest"]], "response_mode": "offline", "answer": "[{\"word\":\"farthest\",\"start\":4.52,\"end\":5.2}]"} {"utterance_id": "4992-23283-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0008.flac", "task_type": "lexical_detection", "query": "proceeded", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"proceeded\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.8, 4.45, "proceeded"]], "response_mode": "offline", "answer": "[{\"word\":\"proceeded\",\"start\":3.8,\"end\":4.45}]"} {"utterance_id": "1221-135767-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.7, 1.12, "pearl"]], "response_mode": "offline", "answer": "[{\"word\":\"pearl\",\"start\":0.7,\"end\":1.12}]"} {"utterance_id": "2300-131720-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0001.flac", "task_type": "semantic_detection", "query": "city", "prompt": "You will hear a complete utterance. When it ends, list every instance of a city name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.9, "london"]], "response_mode": "offline", "answer": "[{\"word\":\"london\",\"start\":0.53,\"end\":0.9}]"} {"utterance_id": "260-123286-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0005.flac", "task_type": "lexical_detection", "query": "wide", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"wide\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.75, 4.38, "wide"]], "response_mode": "offline", "answer": "[{\"word\":\"wide\",\"start\":3.75,\"end\":4.38}]"} {"utterance_id": "1221-135766-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0010.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a person's name (word, start, end). If none, reply [].", "labels": [[9.92, 10.34, "one"], [13.07, 13.43, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":9.92,\"end\":10.34},{\"word\":\"two\",\"start\":13.07,\"end\":13.43}]"} {"utterance_id": "2830-3980-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0004.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.05, 2.45, "christ"]], "response_mode": "offline", "answer": "[{\"word\":\"christ\",\"start\":2.05,\"end\":2.45}]"} {"utterance_id": "4077-13751-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0004.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.76, 2.99, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.76,\"end\":2.99}]"} {"utterance_id": "7176-88083-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0025.flac", "task_type": "lexical_detection", "query": "bank", "prompt": "After the audio ends, extract all occurrences of the word \"bank\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.79, 7.18, "bank"]], "response_mode": "offline", "answer": "[{\"word\":\"bank\",\"start\":6.79,\"end\":7.18}]"} {"utterance_id": "3570-5695-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0009.flac", "task_type": "lexical_detection", "query": "stages", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"stages\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.38, 3.78, "stages"]], "response_mode": "offline", "answer": "[{\"word\":\"stages\",\"start\":3.38,\"end\":3.78}]"} {"utterance_id": "8455-210777-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0017.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.57, 1.18, "wife"]], "response_mode": "offline", "answer": "[{\"word\":\"wife\",\"start\":0.57,\"end\":1.18}]"} {"utterance_id": "1320-122612-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0002.flac", "task_type": "lexical_detection", "query": "proceeding", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"proceeding\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.86, 1.37, "proceeding"]], "response_mode": "offline", "answer": "[{\"word\":\"proceeding\",\"start\":0.86,\"end\":1.37}]"} {"utterance_id": "8555-292519-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0013.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "When the utterance is complete, report all instances of a plant name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.46, 2.94, "plants"]], "response_mode": "offline", "answer": "[{\"word\":\"plants\",\"start\":2.46,\"end\":2.94}]"} {"utterance_id": "5639-40744-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0036.flac", "task_type": "lexical_detection", "query": "destined", "prompt": "After the audio ends, extract all occurrences of the word \"destined\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.67, 3.16, "destined"]], "response_mode": "offline", "answer": "[{\"word\":\"destined\",\"start\":2.67,\"end\":3.16}]"} {"utterance_id": "672-122797-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0022.flac", "task_type": "lexical_detection", "query": "sigh", "prompt": "Process the full audio offline. Find every instance of the word \"sigh\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.96, 6.55, "sigh"]], "response_mode": "offline", "answer": "[{\"word\":\"sigh\",\"start\":5.96,\"end\":6.55}]"} {"utterance_id": "6829-68769-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0032.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[4.13, 4.52, "marshall"]], "response_mode": "offline", "answer": "[{\"word\":\"marshall\",\"start\":4.13,\"end\":4.52}]"} {"utterance_id": "2830-3980-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0039.flac", "task_type": "lexical_detection", "query": "understood", "prompt": "After the audio ends, extract all occurrences of the word \"understood\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.24, 4.93, "understood"]], "response_mode": "offline", "answer": "[{\"word\":\"understood\",\"start\":4.24,\"end\":4.93}]"} {"utterance_id": "3570-5694-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0005.flac", "task_type": "semantic_counting", "query": "food", "prompt": "Once the audio has finished, list each occurrence of a food name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[1.65, 2.25, "one"], [4.06, 4.8, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.65,\"end\":2.25},{\"word\":\"two\",\"start\":4.06,\"end\":4.8}]"} {"utterance_id": "6930-75918-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0004.flac", "task_type": "semantic_detection", "query": "country", "prompt": "After the audio ends, extract all occurrences of a country name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.94, 5.46, "country"]], "response_mode": "offline", "answer": "[{\"word\":\"country\",\"start\":4.94,\"end\":5.46}]"} {"utterance_id": "1089-134686-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0020.flac", "task_type": "lexical_detection", "query": "stolen", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"stolen\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.09, 1.53, "stolen"], [9.86, 10.38, "stolen"]], "response_mode": "offline", "answer": "[{\"word\":\"stolen\",\"start\":1.09,\"end\":1.53},{\"word\":\"stolen\",\"start\":9.86,\"end\":10.38}]"} {"utterance_id": "7176-88083-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0014.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.61, 0.95, "hawk"]], "response_mode": "offline", "answer": "[{\"word\":\"hawk\",\"start\":0.61,\"end\":0.95}]"} {"utterance_id": "6930-76324-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0008.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "When the utterance is complete, report all instances of an ordinal number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.74, 2.07, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":1.74,\"end\":2.07}]"} {"utterance_id": "6930-76324-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0003.flac", "task_type": "lexical_detection", "query": "sense", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"sense\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.83, 1.15, "sense"]], "response_mode": "offline", "answer": "[{\"word\":\"sense\",\"start\":0.83,\"end\":1.15}]"} {"utterance_id": "8455-210777-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0039.flac", "task_type": "lexical_detection", "query": "allowed", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"allowed\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.5, 2.88, "allowed"]], "response_mode": "offline", "answer": "[{\"word\":\"allowed\",\"start\":2.5,\"end\":2.88}]"} {"utterance_id": "1089-134691-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0002.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.22, 4.63, "father's"]], "response_mode": "offline", "answer": "[{\"word\":\"father's\",\"start\":4.22,\"end\":4.63}]"} {"utterance_id": "3575-170457-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0022.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Process the full audio offline. Find every instance of a time of day and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.75, 1.2, "evenings"]], "response_mode": "offline", "answer": "[{\"word\":\"evenings\",\"start\":0.75,\"end\":1.2}]"} {"utterance_id": "1089-134686-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0012.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.7, 1.06, "eyes"], [7.77, 8.31, "flesh"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":0.7,\"end\":1.06},{\"word\":\"flesh\",\"start\":7.77,\"end\":8.31}]"} {"utterance_id": "7729-102255-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0003.flac", "task_type": "semantic_counting", "query": "organization", "prompt": "After the audio ends, tally every instance of an organization name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[6.04, 6.84, "one"], [12.41, 13.03, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.04,\"end\":6.84},{\"word\":\"two\",\"start\":12.41,\"end\":13.03}]"} {"utterance_id": "7729-102255-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0015.flac", "task_type": "semantic_detection", "query": "city", "prompt": "After listening to the whole recording, output every instance of a city name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.07, 6.68, "lawrence"], [13.89, 14.82, "fort leavenworth"]], "response_mode": "offline", "answer": "[{\"word\":\"lawrence\",\"start\":6.07,\"end\":6.68},{\"word\":\"fort leavenworth\",\"start\":13.89,\"end\":14.82}]"} {"utterance_id": "1995-1826-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0012.flac", "task_type": "lexical_detection", "query": "glance", "prompt": "Process the full audio offline. Find every instance of the word \"glance\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.83, 1.27, "glance"]], "response_mode": "offline", "answer": "[{\"word\":\"glance\",\"start\":0.83,\"end\":1.27}]"} {"utterance_id": "237-134500-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0027.flac", "task_type": "lexical_detection", "query": "brooding", "prompt": "After the audio ends, extract all occurrences of the word \"brooding\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.1, 2.47, "brooding"]], "response_mode": "offline", "answer": "[{\"word\":\"brooding\",\"start\":2.1,\"end\":2.47}]"} {"utterance_id": "4077-13754-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0008.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.24, 3.46, "many"], [7.57, 7.94, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"many\",\"start\":3.24,\"end\":3.46},{\"word\":\"one\",\"start\":7.57,\"end\":7.94}]"} {"utterance_id": "61-70970-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0030.flac", "task_type": "lexical_detection", "query": "masters", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"masters\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.89, 1.47, "masters"]], "response_mode": "offline", "answer": "[{\"word\":\"masters\",\"start\":0.89,\"end\":1.47}]"} {"utterance_id": "4446-2275-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0009.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.17, 1.35, "ten"]], "response_mode": "offline", "answer": "[{\"word\":\"ten\",\"start\":1.17,\"end\":1.35}]"} {"utterance_id": "4446-2275-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0002.flac", "task_type": "lexical_detection", "query": "hilda's", "prompt": "Process the full audio offline. Find every instance of the word \"hilda's\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.44, 6.81, "hilda's"]], "response_mode": "offline", "answer": "[{\"word\":\"hilda's\",\"start\":6.44,\"end\":6.81}]"} {"utterance_id": "5683-32866-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0014.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen to the full audio. After it finishes, find every instance of a time of day. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.23, 3.81, "morning"]], "response_mode": "offline", "answer": "[{\"word\":\"morning\",\"start\":3.23,\"end\":3.81}]"} {"utterance_id": "1221-135767-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0001.flac", "task_type": "lexical_detection", "query": "delivery", "prompt": "After listening to the whole recording, output every instance of the word \"delivery\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.1, 3.62, "delivery"]], "response_mode": "offline", "answer": "[{\"word\":\"delivery\",\"start\":3.1,\"end\":3.62}]"} {"utterance_id": "8463-294825-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0015.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.45, 1.95, "one twenty"]], "response_mode": "offline", "answer": "[{\"word\":\"one twenty\",\"start\":1.45,\"end\":1.95}]"} {"utterance_id": "2094-142345-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0048.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.41, 1.5, "captain donnithorne"]], "response_mode": "offline", "answer": "[{\"word\":\"captain donnithorne\",\"start\":0.41,\"end\":1.5}]"} {"utterance_id": "237-126133-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0005.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.2, 2.57, "van"]], "response_mode": "offline", "answer": "[{\"word\":\"van\",\"start\":2.2,\"end\":2.57}]"} {"utterance_id": "2300-131720-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0009.flac", "task_type": "lexical_detection", "query": "alternating", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"alternating\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.06, 4.72, "alternating"]], "response_mode": "offline", "answer": "[{\"word\":\"alternating\",\"start\":4.06,\"end\":4.72}]"} {"utterance_id": "4970-29093-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0019.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.52, 3.86, "philip"]], "response_mode": "offline", "answer": "[{\"word\":\"philip\",\"start\":3.52,\"end\":3.86}]"} {"utterance_id": "237-134493-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0008.flac", "task_type": "lexical_detection", "query": "gathered", "prompt": "After listening to the whole recording, output every instance of the word \"gathered\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.56, 0.94, "gathered"]], "response_mode": "offline", "answer": "[{\"word\":\"gathered\",\"start\":0.56,\"end\":0.94}]"} {"utterance_id": "121-127105-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0028.flac", "task_type": "lexical_detection", "query": "relations", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"relations\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.12, 3.8, "relations"]], "response_mode": "offline", "answer": "[{\"word\":\"relations\",\"start\":3.12,\"end\":3.8}]"} {"utterance_id": "5142-33396-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0023.flac", "task_type": "lexical_detection", "query": "hearth", "prompt": "When the utterance is complete, report all instances of the word \"hearth\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.77, 3.3, "hearth"]], "response_mode": "offline", "answer": "[{\"word\":\"hearth\",\"start\":2.77,\"end\":3.3}]"} {"utterance_id": "7176-92135-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0010.flac", "task_type": "semantic_detection", "query": "food", "prompt": "You will hear a complete utterance. When it ends, list every instance of a food name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.14, 0.48, "ham"]], "response_mode": "offline", "answer": "[{\"word\":\"ham\",\"start\":0.14,\"end\":0.48}]"} {"utterance_id": "5142-33396-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0017.flac", "task_type": "lexical_detection", "query": "ate", "prompt": "Once the audio has finished, identify every instance of the word \"ate\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.37, 0.56, "ate"]], "response_mode": "offline", "answer": "[{\"word\":\"ate\",\"start\":0.37,\"end\":0.56}]"} {"utterance_id": "2300-131720-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0041.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.2, 2.42, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":2.2,\"end\":2.42}]"} {"utterance_id": "7021-79740-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0008.flac", "task_type": "lexical_detection", "query": "piazza", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"piazza\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.89, 5.39, "piazza"], [13.03, 13.61, "piazza"]], "response_mode": "offline", "answer": "[{\"word\":\"piazza\",\"start\":4.89,\"end\":5.39},{\"word\":\"piazza\",\"start\":13.03,\"end\":13.61}]"} {"utterance_id": "5142-33396-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0042.flac", "task_type": "lexical_detection", "query": "tales", "prompt": "Once the audio has finished, identify every instance of the word \"tales\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.57, 1.02, "tales"]], "response_mode": "offline", "answer": "[{\"word\":\"tales\",\"start\":0.57,\"end\":1.02}]"} {"utterance_id": "7729-102255-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0025.flac", "task_type": "lexical_detection", "query": "assumed", "prompt": "When the utterance is complete, report all instances of the word \"assumed\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.48, 0.96, "assumed"]], "response_mode": "offline", "answer": "[{\"word\":\"assumed\",\"start\":0.48,\"end\":0.96}]"} {"utterance_id": "1320-122617-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0005.flac", "task_type": "lexical_detection", "query": "shaggy", "prompt": "When the utterance is complete, report all instances of the word \"shaggy\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.3, 1.74, "shaggy"]], "response_mode": "offline", "answer": "[{\"word\":\"shaggy\",\"start\":1.3,\"end\":1.74}]"} {"utterance_id": "4992-41806-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0012.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After listening to the whole recording, output every instance of a color name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.54, 5.88, "green"]], "response_mode": "offline", "answer": "[{\"word\":\"green\",\"start\":5.54,\"end\":5.88}]"} {"utterance_id": "8463-294825-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0012.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a state or province name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.22, 2.96, "antarctic"]], "response_mode": "offline", "answer": "[{\"word\":\"antarctic\",\"start\":2.22,\"end\":2.96}]"} {"utterance_id": "1580-141084-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0047.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.07, 1.87, "mister soames"]], "response_mode": "offline", "answer": "[{\"word\":\"mister soames\",\"start\":1.07,\"end\":1.87}]"} {"utterance_id": "1320-122617-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0030.flac", "task_type": "lexical_detection", "query": "choose", "prompt": "After listening to the whole recording, output every instance of the word \"choose\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.43, 0.72, "choose"]], "response_mode": "offline", "answer": "[{\"word\":\"choose\",\"start\":0.43,\"end\":0.72}]"} {"utterance_id": "1995-1836-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.59, 1.19, "cresswell"]], "response_mode": "offline", "answer": "[{\"word\":\"cresswell\",\"start\":0.59,\"end\":1.19}]"} {"utterance_id": "237-126133-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.08, 2.64, "jasper"]], "response_mode": "offline", "answer": "[{\"word\":\"jasper\",\"start\":2.08,\"end\":2.64}]"} {"utterance_id": "4507-16021-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0032.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.49, 8.09, "brother"]], "response_mode": "offline", "answer": "[{\"word\":\"brother\",\"start\":7.49,\"end\":8.09}]"} {"utterance_id": "8463-294828-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0038.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.26, 0.84, "thousands"]], "response_mode": "offline", "answer": "[{\"word\":\"thousands\",\"start\":0.26,\"end\":0.84}]"} {"utterance_id": "8555-284447-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0012.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.44, 1.81, "head"]], "response_mode": "offline", "answer": "[{\"word\":\"head\",\"start\":1.44,\"end\":1.81}]"} {"utterance_id": "672-122797-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0025.flac", "task_type": "lexical_detection", "query": "yard", "prompt": "Once the audio has finished, identify every instance of the word \"yard\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.97, 4.39, "yard"]], "response_mode": "offline", "answer": "[{\"word\":\"yard\",\"start\":3.97,\"end\":4.39}]"} {"utterance_id": "260-123440-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0012.flac", "task_type": "lexical_detection", "query": "it'll", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"it'll\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.25, 0.44, "it'll"]], "response_mode": "offline", "answer": "[{\"word\":\"it'll\",\"start\":0.25,\"end\":0.44}]"} {"utterance_id": "672-122797-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0022.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "After the audio ends, tally every instance of a plant name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[2.94, 3.43, "one"], [4.46, 4.88, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.94,\"end\":3.43},{\"word\":\"two\",\"start\":4.46,\"end\":4.88}]"} {"utterance_id": "8455-210777-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0047.flac", "task_type": "lexical_detection", "query": "kidnap", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"kidnap\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.24, 1.69, "kidnap"]], "response_mode": "offline", "answer": "[{\"word\":\"kidnap\",\"start\":1.24,\"end\":1.69}]"} {"utterance_id": "1188-133604-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0023.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "When the utterance is complete, report all instances of an animal name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.63, 5.14, "paroquet"], [7.96, 8.33, "viper"]], "response_mode": "offline", "answer": "[{\"word\":\"paroquet\",\"start\":4.63,\"end\":5.14},{\"word\":\"viper\",\"start\":7.96,\"end\":8.33}]"} {"utterance_id": "672-122797-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0008.flac", "task_type": "lexical_detection", "query": "dragged", "prompt": "Process the full audio offline. Find every instance of the word \"dragged\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[28.38, 29.01, "dragged"]], "response_mode": "offline", "answer": "[{\"word\":\"dragged\",\"start\":28.38,\"end\":29.01}]"} {"utterance_id": "8224-274384-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0011.flac", "task_type": "semantic_detection", "query": "money", "prompt": "You will hear a complete utterance. When it ends, list every instance of a money or currency term you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.58, 10.13, "sum"]], "response_mode": "offline", "answer": "[{\"word\":\"sum\",\"start\":9.58,\"end\":10.13}]"} {"utterance_id": "5683-32865-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0005.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.79, 6.44, "features"]], "response_mode": "offline", "answer": "[{\"word\":\"features\",\"start\":5.79,\"end\":6.44}]"} {"utterance_id": "5142-36377-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0011.flac", "task_type": "lexical_detection", "query": "jago", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"jago\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 0.87, "jago"]], "response_mode": "offline", "answer": "[{\"word\":\"jago\",\"start\":0.52,\"end\":0.87}]"} {"utterance_id": "2094-142345-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0010.flac", "task_type": "lexical_detection", "query": "oak", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"oak\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.2, 11.47, "oak"]], "response_mode": "offline", "answer": "[{\"word\":\"oak\",\"start\":11.2,\"end\":11.47}]"} {"utterance_id": "8224-274381-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0014.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "After listening to the whole recording, output timed count words for a profession (one, two, three, ...) as JSON. If none, reply [].", "labels": [[4.6, 5.23, "one"], [7.23, 7.95, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.6,\"end\":5.23},{\"word\":\"two\",\"start\":7.23,\"end\":7.95}]"} {"utterance_id": "237-134500-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0024.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.24, 1.64, "carl"]], "response_mode": "offline", "answer": "[{\"word\":\"carl\",\"start\":1.24,\"end\":1.64}]"} {"utterance_id": "1580-141084-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0041.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.07, 7.57, "servant"]], "response_mode": "offline", "answer": "[{\"word\":\"servant\",\"start\":7.07,\"end\":7.57}]"} {"utterance_id": "7729-102255-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0046.flac", "task_type": "lexical_detection", "query": "fort", "prompt": "After the audio ends, extract all occurrences of the word \"fort\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.94, 8.4, "fort"]], "response_mode": "offline", "answer": "[{\"word\":\"fort\",\"start\":7.94,\"end\":8.4}]"} {"utterance_id": "1320-122617-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0026.flac", "task_type": "lexical_detection", "query": "main", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"main\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.05, 2.35, "main"]], "response_mode": "offline", "answer": "[{\"word\":\"main\",\"start\":2.05,\"end\":2.35}]"} {"utterance_id": "672-122797-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0013.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an ordinal number with fields word, start, end (seconds). If none, reply [].", "labels": [[5.85, 6.22, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":5.85,\"end\":6.22}]"} {"utterance_id": "672-122797-0064", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0064.flac", "task_type": "lexical_detection", "query": "mice", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"mice\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.79, 2.31, "mice"], [11.32, 11.66, "mice"]], "response_mode": "offline", "answer": "[{\"word\":\"mice\",\"start\":1.79,\"end\":2.31},{\"word\":\"mice\",\"start\":11.32,\"end\":11.66}]"} {"utterance_id": "2094-142345-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0038.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.22, 8.66, "totty"]], "response_mode": "offline", "answer": "[{\"word\":\"totty\",\"start\":8.22,\"end\":8.66}]"} {"utterance_id": "7021-85628-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0011.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[8.24, 8.77, "side"]], "response_mode": "offline", "answer": "[{\"word\":\"side\",\"start\":8.24,\"end\":8.77}]"} {"utterance_id": "8224-274384-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0003.flac", "task_type": "lexical_detection", "query": "gift", "prompt": "Process the full audio offline. Find every instance of the word \"gift\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.75, 3.37, "gift"]], "response_mode": "offline", "answer": "[{\"word\":\"gift\",\"start\":2.75,\"end\":3.37}]"} {"utterance_id": "4446-2271-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0012.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.62, 1.19, "sir harry"]], "response_mode": "offline", "answer": "[{\"word\":\"sir harry\",\"start\":0.62,\"end\":1.19}]"} {"utterance_id": "5683-32879-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0000.flac", "task_type": "lexical_detection", "query": "garden", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"garden\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.74, 7.14, "garden"]], "response_mode": "offline", "answer": "[{\"word\":\"garden\",\"start\":6.74,\"end\":7.14}]"} {"utterance_id": "237-126133-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0017.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.27, 5.74, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":5.27,\"end\":5.74}]"} {"utterance_id": "672-122797-0058", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0058.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.6, 2.82, "humpy dumpy"]], "response_mode": "offline", "answer": "[{\"word\":\"humpy dumpy\",\"start\":1.6,\"end\":2.82}]"} {"utterance_id": "4077-13751-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0021.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.43, 8.31, "governor boggs"]], "response_mode": "offline", "answer": "[{\"word\":\"governor boggs\",\"start\":7.43,\"end\":8.31}]"} {"utterance_id": "4446-2273-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0025.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.62, 2.9, "marie"]], "response_mode": "offline", "answer": "[{\"word\":\"marie\",\"start\":2.62,\"end\":2.9}]"} {"utterance_id": "5142-36377-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0024.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.14, 4.57, "farmers"]], "response_mode": "offline", "answer": "[{\"word\":\"farmers\",\"start\":4.14,\"end\":4.57}]"} {"utterance_id": "61-70968-0048", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0048.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Process the full audio offline. Find every instance of a country name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.33, 1.7, "england"]], "response_mode": "offline", "answer": "[{\"word\":\"england\",\"start\":1.33,\"end\":1.7}]"} {"utterance_id": "4077-13754-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0013.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[11.35, 11.69, "many"]], "response_mode": "offline", "answer": "[{\"word\":\"many\",\"start\":11.35,\"end\":11.69}]"} {"utterance_id": "8463-294825-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0003.flac", "task_type": "lexical_detection", "query": "nautilus", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"nautilus\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.45, 4.07, "nautilus"]], "response_mode": "offline", "answer": "[{\"word\":\"nautilus\",\"start\":3.45,\"end\":4.07}]"} {"utterance_id": "8230-279154-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0014.flac", "task_type": "lexical_detection", "query": "images", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"images\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.54, 4.05, "images"]], "response_mode": "offline", "answer": "[{\"word\":\"images\",\"start\":3.54,\"end\":4.05}]"} {"utterance_id": "7021-85628-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0007.flac", "task_type": "lexical_detection", "query": "marble", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"marble\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.69, 5.11, "marble"]], "response_mode": "offline", "answer": "[{\"word\":\"marble\",\"start\":4.69,\"end\":5.11}]"} {"utterance_id": "5142-33396-0058", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0058.flac", "task_type": "lexical_detection", "query": "king", "prompt": "Once the audio has finished, identify every instance of the word \"king\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.59, 2.02, "king"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":1.59,\"end\":2.02}]"} {"utterance_id": "7176-92135-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0039.flac", "task_type": "semantic_detection", "query": "food", "prompt": "When the utterance is complete, report all instances of a food name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.25, 0.61, "tea"]], "response_mode": "offline", "answer": "[{\"word\":\"tea\",\"start\":0.25,\"end\":0.61}]"} {"utterance_id": "121-127105-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0005.flac", "task_type": "lexical_detection", "query": "finds", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"finds\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.06, 5.41, "finds"]], "response_mode": "offline", "answer": "[{\"word\":\"finds\",\"start\":5.06,\"end\":5.41}]"} {"utterance_id": "5142-36377-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0018.flac", "task_type": "lexical_detection", "query": "naomi", "prompt": "After the audio ends, extract all occurrences of the word \"naomi\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.84, 1.26, "naomi"]], "response_mode": "offline", "answer": "[{\"word\":\"naomi\",\"start\":0.84,\"end\":1.26}]"} {"utterance_id": "7176-88083-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0000.flac", "task_type": "lexical_detection", "query": "tumult", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"tumult\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.35, 1.82, "tumult"]], "response_mode": "offline", "answer": "[{\"word\":\"tumult\",\"start\":1.35,\"end\":1.82}]"} {"utterance_id": "7021-85628-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0015.flac", "task_type": "lexical_detection", "query": "pockets", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"pockets\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.98, 4.44, "pockets"]], "response_mode": "offline", "answer": "[{\"word\":\"pockets\",\"start\":3.98,\"end\":4.44}]"} {"utterance_id": "3570-5695-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0015.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.04, 7.53, "workmen"]], "response_mode": "offline", "answer": "[{\"word\":\"workmen\",\"start\":7.04,\"end\":7.53}]"} {"utterance_id": "121-127105-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0024.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.89, 6.18, "third"], [13.78, 14.24, "fourth"]], "response_mode": "offline", "answer": "[{\"word\":\"third\",\"start\":5.89,\"end\":6.18},{\"word\":\"fourth\",\"start\":13.78,\"end\":14.24}]"} {"utterance_id": "672-122797-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0035.flac", "task_type": "lexical_detection", "query": "tell", "prompt": "After the audio ends, extract all occurrences of the word \"tell\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.94, 1.34, "tell"]], "response_mode": "offline", "answer": "[{\"word\":\"tell\",\"start\":0.94,\"end\":1.34}]"} {"utterance_id": "672-122797-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0012.flac", "task_type": "lexical_detection", "query": "glorious", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"glorious\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.87, 4.64, "glorious"]], "response_mode": "offline", "answer": "[{\"word\":\"glorious\",\"start\":3.87,\"end\":4.64}]"} {"utterance_id": "1580-141084-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0023.flac", "task_type": "lexical_detection", "query": "allowing", "prompt": "After listening to the whole recording, output every instance of the word \"allowing\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.85, 6.22, "allowing"]], "response_mode": "offline", "answer": "[{\"word\":\"allowing\",\"start\":5.85,\"end\":6.22}]"} {"utterance_id": "7729-102255-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0035.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Once the audio has finished, identify every instance of a city name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.97, 5.36, "town"]], "response_mode": "offline", "answer": "[{\"word\":\"town\",\"start\":4.97,\"end\":5.36}]"} {"utterance_id": "121-127105-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0000.flac", "task_type": "lexical_detection", "query": "observation", "prompt": "After the audio ends, extract all occurrences of the word \"observation\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.03, 1.77, "observation"]], "response_mode": "offline", "answer": "[{\"word\":\"observation\",\"start\":1.03,\"end\":1.77}]"} {"utterance_id": "7127-75946-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0012.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "After listening to the whole recording, output every instance of a weather term with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.33, 3.0, "lightning"]], "response_mode": "offline", "answer": "[{\"word\":\"lightning\",\"start\":2.33,\"end\":3.0}]"} {"utterance_id": "7127-75947-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0037.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.47, 3.11, "montalais"], [7.88, 8.34, "louise"]], "response_mode": "offline", "answer": "[{\"word\":\"montalais\",\"start\":2.47,\"end\":3.11},{\"word\":\"louise\",\"start\":7.88,\"end\":8.34}]"} {"utterance_id": "5142-36377-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0014.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Once the audio has finished, identify every instance of a color name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[9.88, 10.19, "gray"]], "response_mode": "offline", "answer": "[{\"word\":\"gray\",\"start\":9.88,\"end\":10.19}]"} {"utterance_id": "4446-2275-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0024.flac", "task_type": "lexical_detection", "query": "gratitude", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"gratitude\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.55, 2.13, "gratitude"]], "response_mode": "offline", "answer": "[{\"word\":\"gratitude\",\"start\":1.55,\"end\":2.13}]"} {"utterance_id": "5142-33396-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0022.flac", "task_type": "lexical_detection", "query": "viking", "prompt": "Once the audio has finished, identify every instance of the word \"viking\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.18, 3.62, "viking"]], "response_mode": "offline", "answer": "[{\"word\":\"viking\",\"start\":3.18,\"end\":3.62}]"} {"utterance_id": "121-127105-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0021.flac", "task_type": "lexical_detection", "query": "douglas", "prompt": "When the utterance is complete, report all instances of the word \"douglas\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.31, 1.88, "douglas"]], "response_mode": "offline", "answer": "[{\"word\":\"douglas\",\"start\":1.31,\"end\":1.88}]"} {"utterance_id": "5639-40744-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0020.flac", "task_type": "lexical_detection", "query": "embracing", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"embracing\" with fields word, start, end (seconds). If none, reply [].", "labels": [[5.54, 6.11, "embracing"]], "response_mode": "offline", "answer": "[{\"word\":\"embracing\",\"start\":5.54,\"end\":6.11}]"} {"utterance_id": "5683-32879-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0018.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "Listen carefully to the entire clip, then report every instance of a disease name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.41, 6.07, "madness"]], "response_mode": "offline", "answer": "[{\"word\":\"madness\",\"start\":5.41,\"end\":6.07}]"} {"utterance_id": "3570-5696-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0003.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an organization name with fields word, start, end (seconds). If none, reply [].", "labels": [[10.82, 11.6, "organizations"]], "response_mode": "offline", "answer": "[{\"word\":\"organizations\",\"start\":10.82,\"end\":11.6}]"} {"utterance_id": "672-122797-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0039.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.16, 3.8, "fir tree"]], "response_mode": "offline", "answer": "[{\"word\":\"fir tree\",\"start\":3.16,\"end\":3.8}]"} {"utterance_id": "4507-16021-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0035.flac", "task_type": "lexical_detection", "query": "history", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"history\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.78, 1.38, "history"]], "response_mode": "offline", "answer": "[{\"word\":\"history\",\"start\":0.78,\"end\":1.38}]"} {"utterance_id": "7729-102255-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0043.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.13, 1.63, "officer"]], "response_mode": "offline", "answer": "[{\"word\":\"officer\",\"start\":1.13,\"end\":1.63}]"} {"utterance_id": "61-70968-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0046.flac", "task_type": "lexical_detection", "query": "demanded", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"demanded\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.81, 3.37, "demanded"]], "response_mode": "offline", "answer": "[{\"word\":\"demanded\",\"start\":2.81,\"end\":3.37}]"} {"utterance_id": "908-157963-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0011.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Once the audio has finished, identify every instance of an animal name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.45, 7.08, "cow"], [9.66, 10.26, "steed"]], "response_mode": "offline", "answer": "[{\"word\":\"cow\",\"start\":6.45,\"end\":7.08},{\"word\":\"steed\",\"start\":9.66,\"end\":10.26}]"} {"utterance_id": "1284-1180-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0017.flac", "task_type": "semantic_detection", "query": "color", "prompt": "You will hear a complete utterance. When it ends, list every instance of a color name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.67, 2.92, "blue"], [4.25, 4.55, "blue"]], "response_mode": "offline", "answer": "[{\"word\":\"blue\",\"start\":2.67,\"end\":2.92},{\"word\":\"blue\",\"start\":4.25,\"end\":4.55}]"} {"utterance_id": "2830-3979-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0003.flac", "task_type": "lexical_detection", "query": "undertaking", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"undertaking\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.75, 1.39, "undertaking"]], "response_mode": "offline", "answer": "[{\"word\":\"undertaking\",\"start\":0.75,\"end\":1.39}]"} {"utterance_id": "260-123440-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0018.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an animal name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.47, 3.15, "mouse"]], "response_mode": "offline", "answer": "[{\"word\":\"mouse\",\"start\":2.47,\"end\":3.15}]"} {"utterance_id": "2830-3980-0065", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0065.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen carefully to the entire clip, then report every instance of a family relation as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.96, 4.22, "son"]], "response_mode": "offline", "answer": "[{\"word\":\"son\",\"start\":3.96,\"end\":4.22}]"} {"utterance_id": "5683-32865-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0003.flac", "task_type": "semantic_counting", "query": "kinship", "prompt": "After the audio ends, tally every instance of a family relation. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.66, 1.22, "one"], [2.61, 3.33, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.66,\"end\":1.22},{\"word\":\"two\",\"start\":2.61,\"end\":3.33}]"} {"utterance_id": "3729-6852-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0020.flac", "task_type": "semantic_detection", "query": "city", "prompt": "After listening to the whole recording, output every instance of a city name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.42, 8.93, "city"]], "response_mode": "offline", "answer": "[{\"word\":\"city\",\"start\":8.42,\"end\":8.93}]"} {"utterance_id": "1089-134691-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0012.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.49, 9.89, "beggar's"]], "response_mode": "offline", "answer": "[{\"word\":\"beggar's\",\"start\":9.49,\"end\":9.89}]"} {"utterance_id": "5639-40744-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0031.flac", "task_type": "lexical_detection", "query": "summer's", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"summer's\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[10.89, 11.29, "summer's"]], "response_mode": "offline", "answer": "[{\"word\":\"summer's\",\"start\":10.89,\"end\":11.29}]"} {"utterance_id": "1188-133604-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0044.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an ordinal number with fields word, start, end (seconds). If none, reply [].", "labels": [[6.98, 7.35, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":6.98,\"end\":7.35}]"} {"utterance_id": "2961-960-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0006.flac", "task_type": "lexical_detection", "query": "fancies", "prompt": "Process the full audio offline. Find every instance of the word \"fancies\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.57, 1.15, "fancies"]], "response_mode": "offline", "answer": "[{\"word\":\"fancies\",\"start\":0.57,\"end\":1.15}]"} {"utterance_id": "260-123288-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0008.flac", "task_type": "lexical_detection", "query": "storm", "prompt": "After the audio ends, extract all occurrences of the word \"storm\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.18, 1.71, "storm"]], "response_mode": "offline", "answer": "[{\"word\":\"storm\",\"start\":1.18,\"end\":1.71}]"} {"utterance_id": "237-126133-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0010.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.84, 5.16, "mammy"]], "response_mode": "offline", "answer": "[{\"word\":\"mammy\",\"start\":4.84,\"end\":5.16}]"} {"utterance_id": "2094-142345-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0022.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 1.36, "mister ottley's"]], "response_mode": "offline", "answer": "[{\"word\":\"mister ottley's\",\"start\":0.51,\"end\":1.36}]"} {"utterance_id": "6829-68769-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0027.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[2.29, 2.7, "kenneth"]], "response_mode": "offline", "answer": "[{\"word\":\"kenneth\",\"start\":2.29,\"end\":2.7}]"} {"utterance_id": "908-31957-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0014.flac", "task_type": "lexical_detection", "query": "kiss", "prompt": "Process the full audio offline. Find every instance of the word \"kiss\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.73, 7.3, "kiss"]], "response_mode": "offline", "answer": "[{\"word\":\"kiss\",\"start\":6.73,\"end\":7.3}]"} {"utterance_id": "8463-294825-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0004.flac", "task_type": "lexical_detection", "query": "difficult", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"difficult\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.77, 2.44, "difficult"]], "response_mode": "offline", "answer": "[{\"word\":\"difficult\",\"start\":1.77,\"end\":2.44}]"} {"utterance_id": "260-123286-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0002.flac", "task_type": "lexical_detection", "query": "sufferings", "prompt": "Process the full audio offline. Find every instance of the word \"sufferings\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.91, 2.61, "sufferings"]], "response_mode": "offline", "answer": "[{\"word\":\"sufferings\",\"start\":1.91,\"end\":2.61}]"} {"utterance_id": "5105-28241-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0016.flac", "task_type": "lexical_detection", "query": "imagined", "prompt": "After listening to the whole recording, output every instance of the word \"imagined\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.27, 5.86, "imagined"]], "response_mode": "offline", "answer": "[{\"word\":\"imagined\",\"start\":5.27,\"end\":5.86}]"} {"utterance_id": "6829-68769-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0008.flac", "task_type": "lexical_detection", "query": "fairview", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"fairview\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.2, 0.62, "fairview"]], "response_mode": "offline", "answer": "[{\"word\":\"fairview\",\"start\":0.2,\"end\":0.62}]"} {"utterance_id": "8230-279154-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0028.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen to the full audio. After it finishes, find every instance of a food name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.1, 6.64, "breakfast"]], "response_mode": "offline", "answer": "[{\"word\":\"breakfast\",\"start\":6.1,\"end\":6.64}]"} {"utterance_id": "2094-142345-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0008.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "Listen to the full audio. Count every occurrence of an animal name. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[16.85, 17.08, "one"], [25.47, 25.97, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":16.85,\"end\":17.08},{\"word\":\"two\",\"start\":25.47,\"end\":25.97}]"} {"utterance_id": "6829-68769-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0037.flac", "task_type": "lexical_detection", "query": "lots", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"lots\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.86, 1.14, "lots"]], "response_mode": "offline", "answer": "[{\"word\":\"lots\",\"start\":0.86,\"end\":1.14}]"} {"utterance_id": "1320-122612-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0003.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen to the full audio. After it finishes, find every instance of a plant name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.11, 2.78, "trees"]], "response_mode": "offline", "answer": "[{\"word\":\"trees\",\"start\":2.11,\"end\":2.78}]"} {"utterance_id": "4992-41806-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0000.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.17, 3.6, "several"]], "response_mode": "offline", "answer": "[{\"word\":\"several\",\"start\":3.17,\"end\":3.6}]"} {"utterance_id": "1188-133604-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0025.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.35, 5.79, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":5.35,\"end\":5.79}]"} {"utterance_id": "7729-102255-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0014.flac", "task_type": "lexical_detection", "query": "conspiracy", "prompt": "After listening to the whole recording, output every instance of the word \"conspiracy\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.17, 1.99, "conspiracy"]], "response_mode": "offline", "answer": "[{\"word\":\"conspiracy\",\"start\":1.17,\"end\":1.99}]"} {"utterance_id": "2961-961-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0008.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[6.12, 6.8, "actors"]], "response_mode": "offline", "answer": "[{\"word\":\"actors\",\"start\":6.12,\"end\":6.8}]"} {"utterance_id": "2961-961-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0001.flac", "task_type": "lexical_detection", "query": "ideal", "prompt": "Process the full audio offline. Find every instance of the word \"ideal\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.51, 3.0, "ideal"]], "response_mode": "offline", "answer": "[{\"word\":\"ideal\",\"start\":2.51,\"end\":3.0}]"} {"utterance_id": "3729-6852-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0007.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 0.73, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":0.51,\"end\":0.73}]"} {"utterance_id": "1188-133604-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0025.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "You will hear a complete utterance. When it ends, list every instance of a weather term you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.89, 6.29, "cloud"]], "response_mode": "offline", "answer": "[{\"word\":\"cloud\",\"start\":5.89,\"end\":6.29}]"} {"utterance_id": "260-123286-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0004.flac", "task_type": "lexical_detection", "query": "reason", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"reason\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.05, 2.4, "reason"]], "response_mode": "offline", "answer": "[{\"word\":\"reason\",\"start\":2.05,\"end\":2.4}]"} {"utterance_id": "1995-1837-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0029.flac", "task_type": "lexical_detection", "query": "darted", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"darted\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.38, 0.75, "darted"]], "response_mode": "offline", "answer": "[{\"word\":\"darted\",\"start\":0.38,\"end\":0.75}]"} {"utterance_id": "1580-141084-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0014.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.31, 1.96, "servant"]], "response_mode": "offline", "answer": "[{\"word\":\"servant\",\"start\":1.31,\"end\":1.96}]"} {"utterance_id": "260-123286-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0018.flac", "task_type": "lexical_detection", "query": "creatures", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"creatures\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.42, 3.95, "creatures"]], "response_mode": "offline", "answer": "[{\"word\":\"creatures\",\"start\":3.42,\"end\":3.95}]"} {"utterance_id": "2300-131720-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0016.flac", "task_type": "lexical_counting", "query": "engine", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"engine\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[3.23, 3.78, "one"], [9.07, 9.68, "two"], [15.74, 16.15, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.23,\"end\":3.78},{\"word\":\"two\",\"start\":9.07,\"end\":9.68},{\"word\":\"three\",\"start\":15.74,\"end\":16.15}]"} {"utterance_id": "3729-6852-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0035.flac", "task_type": "lexical_counting", "query": "affection", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"affection\" (word, start, end). If none, reply [].", "labels": [[12.32, 12.98, "one"], [16.64, 17.12, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":12.32,\"end\":12.98},{\"word\":\"two\",\"start\":16.64,\"end\":17.12}]"} {"utterance_id": "8455-210777-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0030.flac", "task_type": "lexical_detection", "query": "unpleasant", "prompt": "When the utterance is complete, report all instances of the word \"unpleasant\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.45, 11.99, "unpleasant"]], "response_mode": "offline", "answer": "[{\"word\":\"unpleasant\",\"start\":11.45,\"end\":11.99}]"} {"utterance_id": "4446-2273-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0019.flac", "task_type": "lexical_detection", "query": "smoke", "prompt": "After listening to the whole recording, output every instance of the word \"smoke\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.58, 3.03, "smoke"]], "response_mode": "offline", "answer": "[{\"word\":\"smoke\",\"start\":2.58,\"end\":3.03}]"} {"utterance_id": "5639-40744-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0039.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.72, 4.2, "rodolfo"]], "response_mode": "offline", "answer": "[{\"word\":\"rodolfo\",\"start\":3.72,\"end\":4.2}]"} {"utterance_id": "908-31957-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0009.flac", "task_type": "lexical_detection", "query": "grown", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"grown\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.48, 1.84, "grown"]], "response_mode": "offline", "answer": "[{\"word\":\"grown\",\"start\":1.48,\"end\":1.84}]"} {"utterance_id": "4507-16021-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0007.flac", "task_type": "lexical_detection", "query": "shudder", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"shudder\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.53, 2.18, "shudder"]], "response_mode": "offline", "answer": "[{\"word\":\"shudder\",\"start\":1.53,\"end\":2.18}]"} {"utterance_id": "8230-279154-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0002.flac", "task_type": "lexical_counting", "query": "seem", "prompt": "Listen to the entire clip, then report the count sequence for the word \"seem\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.54, 8.75, "one"], [10.41, 10.65, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":8.54,\"end\":8.75},{\"word\":\"two\",\"start\":10.41,\"end\":10.65}]"} {"utterance_id": "2094-142345-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0052.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen carefully to the entire clip, then report every instance of an animal name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.53, 6.26, "horses"]], "response_mode": "offline", "answer": "[{\"word\":\"horses\",\"start\":5.53,\"end\":6.26}]"} {"utterance_id": "4507-16021-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0002.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.29, 2.79, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":2.29,\"end\":2.79}]"} {"utterance_id": "260-123286-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0003.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.37, 5.71, "eye"]], "response_mode": "offline", "answer": "[{\"word\":\"eye\",\"start\":5.37,\"end\":5.71}]"} {"utterance_id": "1995-1826-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0016.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Do not answer until the audio ends. Then list each instance of a plant name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.76, 2.18, "leaves"]], "response_mode": "offline", "answer": "[{\"word\":\"leaves\",\"start\":1.76,\"end\":2.18}]"} {"utterance_id": "7127-75947-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0021.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[1.55, 2.24, "poetess"]], "response_mode": "offline", "answer": "[{\"word\":\"poetess\",\"start\":1.55,\"end\":2.24}]"} {"utterance_id": "7176-92135-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0038.flac", "task_type": "lexical_detection", "query": "charles", "prompt": "After the audio ends, extract all occurrences of the word \"charles\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.81, 5.18, "charles"]], "response_mode": "offline", "answer": "[{\"word\":\"charles\",\"start\":4.81,\"end\":5.18}]"} {"utterance_id": "1284-1180-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0024.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.73, 1.01, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.73,\"end\":1.01}]"} {"utterance_id": "4507-16021-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0000.flac", "task_type": "lexical_detection", "query": "chapter", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"chapter\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.21, 0.64, "chapter"]], "response_mode": "offline", "answer": "[{\"word\":\"chapter\",\"start\":0.21,\"end\":0.64}]"} {"utterance_id": "2300-131720-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0032.flac", "task_type": "lexical_detection", "query": "standard", "prompt": "After listening to the whole recording, output every instance of the word \"standard\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.73, 1.32, "standard"]], "response_mode": "offline", "answer": "[{\"word\":\"standard\",\"start\":0.73,\"end\":1.32}]"} {"utterance_id": "8230-279154-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0029.flac", "task_type": "lexical_detection", "query": "previous", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"previous\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.71, 5.35, "previous"]], "response_mode": "offline", "answer": "[{\"word\":\"previous\",\"start\":4.71,\"end\":5.35}]"} {"utterance_id": "2094-142345-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0051.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the full audio. After it finishes, find every instance of a profession. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.38, 4.9, "shepherd"]], "response_mode": "offline", "answer": "[{\"word\":\"shepherd\",\"start\":4.38,\"end\":4.9}]"} {"utterance_id": "8224-274384-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0011.flac", "task_type": "lexical_detection", "query": "previously", "prompt": "After listening to the whole recording, output every instance of the word \"previously\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.9, 4.57, "previously"]], "response_mode": "offline", "answer": "[{\"word\":\"previously\",\"start\":3.9,\"end\":4.57}]"} {"utterance_id": "2094-142345-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0010.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen carefully to the entire clip, then report every instance of a food name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[25.53, 25.8, "dinner"]], "response_mode": "offline", "answer": "[{\"word\":\"dinner\",\"start\":25.53,\"end\":25.8}]"} {"utterance_id": "7127-75946-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0004.flac", "task_type": "semantic_detection", "query": "money", "prompt": "After the audio ends, extract all occurrences of a money or currency term. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.47, 2.78, "money"]], "response_mode": "offline", "answer": "[{\"word\":\"money\",\"start\":2.47,\"end\":2.78}]"} {"utterance_id": "4446-2275-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0006.flac", "task_type": "lexical_detection", "query": "happening", "prompt": "When the utterance is complete, report all instances of the word \"happening\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.58, 3.02, "happening"]], "response_mode": "offline", "answer": "[{\"word\":\"happening\",\"start\":2.58,\"end\":3.02}]"} {"utterance_id": "2094-142345-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0025.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.84, 3.36, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":2.84,\"end\":3.36}]"} {"utterance_id": "260-123440-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0015.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.61, 3.19, "alice"]], "response_mode": "offline", "answer": "[{\"word\":\"alice\",\"start\":2.61,\"end\":3.19}]"} {"utterance_id": "7176-88083-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0010.flac", "task_type": "lexical_detection", "query": "growled", "prompt": "After listening to the whole recording, output every instance of the word \"growled\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.7, 1.09, "growled"]], "response_mode": "offline", "answer": "[{\"word\":\"growled\",\"start\":0.7,\"end\":1.09}]"} {"utterance_id": "2094-142345-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0028.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "When the utterance is complete, report how a person's name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.0, 2.36, "one"], [3.9, 4.86, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.0,\"end\":2.36},{\"word\":\"two\",\"start\":3.9,\"end\":4.86}]"} {"utterance_id": "2830-3980-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0001.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.37, 3.77, "paul"]], "response_mode": "offline", "answer": "[{\"word\":\"paul\",\"start\":3.37,\"end\":3.77}]"} {"utterance_id": "2094-142345-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0000.flac", "task_type": "semantic_detection", "query": "color", "prompt": "You will hear a complete utterance. When it ends, list every instance of a color name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.38, 2.66, "red"], [4.6, 5.24, "pale"], [12.48, 12.81, "red"]], "response_mode": "offline", "answer": "[{\"word\":\"red\",\"start\":2.38,\"end\":2.66},{\"word\":\"pale\",\"start\":4.6,\"end\":5.24},{\"word\":\"red\",\"start\":12.48,\"end\":12.81}]"} {"utterance_id": "7176-88083-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0012.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an animal name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.35, 0.63, "hawk"]], "response_mode": "offline", "answer": "[{\"word\":\"hawk\",\"start\":0.35,\"end\":0.63}]"} {"utterance_id": "672-122797-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0003.flac", "task_type": "lexical_detection", "query": "bear", "prompt": "Once the audio has finished, identify every instance of the word \"bear\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.33, 3.68, "bear"]], "response_mode": "offline", "answer": "[{\"word\":\"bear\",\"start\":3.33,\"end\":3.68}]"} {"utterance_id": "4507-16021-0057", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0057.flac", "task_type": "lexical_detection", "query": "suffer", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"suffer\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.8, 1.3, "suffer"]], "response_mode": "offline", "answer": "[{\"word\":\"suffer\",\"start\":0.8,\"end\":1.3}]"} {"utterance_id": "4446-2275-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0022.flac", "task_type": "lexical_detection", "query": "summer", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"summer\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.32, 2.83, "summer"]], "response_mode": "offline", "answer": "[{\"word\":\"summer\",\"start\":2.32,\"end\":2.83}]"} {"utterance_id": "5683-32879-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0016.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.57, 7.08, "dorcas"]], "response_mode": "offline", "answer": "[{\"word\":\"dorcas\",\"start\":6.57,\"end\":7.08}]"} {"utterance_id": "2300-131720-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0017.flac", "task_type": "lexical_counting", "query": "system", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"system\" (word, start, end). If none, reply [].", "labels": [[4.23, 4.68, "one"], [6.92, 7.5, "two"], [16.36, 16.89, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.23,\"end\":4.68},{\"word\":\"two\",\"start\":6.92,\"end\":7.5},{\"word\":\"three\",\"start\":16.36,\"end\":16.89}]"} {"utterance_id": "3570-5694-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0015.flac", "task_type": "lexical_detection", "query": "classed", "prompt": "After listening to the whole recording, output every instance of the word \"classed\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.12, 5.56, "classed"]], "response_mode": "offline", "answer": "[{\"word\":\"classed\",\"start\":5.12,\"end\":5.56}]"} {"utterance_id": "5683-32866-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0002.flac", "task_type": "lexical_detection", "query": "wise", "prompt": "Once the audio has finished, identify every instance of the word \"wise\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.31, 1.76, "wise"]], "response_mode": "offline", "answer": "[{\"word\":\"wise\",\"start\":1.31,\"end\":1.76}]"} {"utterance_id": "2961-960-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0010.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.3, 14.89, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":14.3,\"end\":14.89}]"} {"utterance_id": "7021-85628-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0023.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.07, 1.43, "brother"]], "response_mode": "offline", "answer": "[{\"word\":\"brother\",\"start\":1.07,\"end\":1.43}]"} {"utterance_id": "5142-33396-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0033.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen carefully to the entire clip, then report every instance of a food name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.1, 3.46, "ale"]], "response_mode": "offline", "answer": "[{\"word\":\"ale\",\"start\":3.1,\"end\":3.46}]"} {"utterance_id": "1320-122617-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0017.flac", "task_type": "lexical_detection", "query": "safety", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"safety\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.83, 2.34, "safety"]], "response_mode": "offline", "answer": "[{\"word\":\"safety\",\"start\":1.83,\"end\":2.34}]"} {"utterance_id": "2094-142345-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0018.flac", "task_type": "lexical_detection", "query": "taught", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"taught\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.35, 0.57, "taught"]], "response_mode": "offline", "answer": "[{\"word\":\"taught\",\"start\":0.35,\"end\":0.57}]"} {"utterance_id": "8230-279154-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0025.flac", "task_type": "lexical_counting", "query": "memory", "prompt": "Listen to the entire clip, then report the count sequence for the word \"memory\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.43, 4.99, "one"], [18.77, 19.2, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.43,\"end\":4.99},{\"word\":\"two\",\"start\":18.77,\"end\":19.2}]"} {"utterance_id": "4970-29095-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0020.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.28, 3.54, "ruth"]], "response_mode": "offline", "answer": "[{\"word\":\"ruth\",\"start\":3.28,\"end\":3.54}]"} {"utterance_id": "2300-131720-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0016.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Process the full audio offline. Find every instance of a city name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.29, 11.88, "new york"]], "response_mode": "offline", "answer": "[{\"word\":\"new york\",\"start\":11.29,\"end\":11.88}]"} {"utterance_id": "2961-960-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0003.flac", "task_type": "lexical_detection", "query": "connect", "prompt": "Once the audio has finished, identify every instance of the word \"connect\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[12.41, 12.9, "connect"]], "response_mode": "offline", "answer": "[{\"word\":\"connect\",\"start\":12.41,\"end\":12.9}]"} {"utterance_id": "4077-13751-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.12, 7.59, "mormon"]], "response_mode": "offline", "answer": "[{\"word\":\"mormon\",\"start\":7.12,\"end\":7.59}]"} {"utterance_id": "7176-92135-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0026.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.76, 1.24, "hamlet"]], "response_mode": "offline", "answer": "[{\"word\":\"hamlet\",\"start\":0.76,\"end\":1.24}]"} {"utterance_id": "1188-133604-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0033.flac", "task_type": "lexical_detection", "query": "formally", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"formally\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.06, 2.65, "formally"]], "response_mode": "offline", "answer": "[{\"word\":\"formally\",\"start\":2.06,\"end\":2.65}]"} {"utterance_id": "5142-33396-0055", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0055.flac", "task_type": "lexical_detection", "query": "pointed", "prompt": "When the utterance is complete, report all instances of the word \"pointed\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.8, 1.15, "pointed"]], "response_mode": "offline", "answer": "[{\"word\":\"pointed\",\"start\":0.8,\"end\":1.15}]"} {"utterance_id": "3575-170457-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0039.flac", "task_type": "semantic_detection", "query": "date", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a date with fields word, start, end (seconds). If none, reply [].", "labels": [[0.61, 0.99, "christmas"]], "response_mode": "offline", "answer": "[{\"word\":\"christmas\",\"start\":0.61,\"end\":0.99}]"} {"utterance_id": "1284-1180-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0002.flac", "task_type": "semantic_counting", "query": "color", "prompt": "Listen to the entire clip, then report the count sequence for a color name as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.7, 4.97, "one"], [6.42, 6.77, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.7,\"end\":4.97},{\"word\":\"two\",\"start\":6.42,\"end\":6.77}]"} {"utterance_id": "3729-6852-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0045.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[16.27, 16.68, "twenty"]], "response_mode": "offline", "answer": "[{\"word\":\"twenty\",\"start\":16.27,\"end\":16.68}]"} {"utterance_id": "6930-76324-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0023.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Do not answer until the audio ends. Then list each instance of a money or currency term as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.86, 1.13, "money"]], "response_mode": "offline", "answer": "[{\"word\":\"money\",\"start\":0.86,\"end\":1.13}]"} {"utterance_id": "2961-961-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0022.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[6.55, 7.77, "natural philosopher"]], "response_mode": "offline", "answer": "[{\"word\":\"natural philosopher\",\"start\":6.55,\"end\":7.77}]"} {"utterance_id": "4446-2273-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0016.flac", "task_type": "semantic_detection", "query": "color", "prompt": "When the utterance is complete, report all instances of a color name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.66, 3.02, "blue"]], "response_mode": "offline", "answer": "[{\"word\":\"blue\",\"start\":2.66,\"end\":3.02}]"} {"utterance_id": "4992-41797-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0011.flac", "task_type": "lexical_detection", "query": "straightway", "prompt": "When the utterance is complete, report all instances of the word \"straightway\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.55, 3.22, "straightway"]], "response_mode": "offline", "answer": "[{\"word\":\"straightway\",\"start\":2.55,\"end\":3.22}]"} {"utterance_id": "4992-23283-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0019.flac", "task_type": "lexical_detection", "query": "lord", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"lord\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.84, 6.16, "lord"]], "response_mode": "offline", "answer": "[{\"word\":\"lord\",\"start\":5.84,\"end\":6.16}]"} {"utterance_id": "4970-29095-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0037.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.97, 5.24, "ruth"]], "response_mode": "offline", "answer": "[{\"word\":\"ruth\",\"start\":4.97,\"end\":5.24}]"} {"utterance_id": "8224-274381-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0012.flac", "task_type": "lexical_detection", "query": "lined", "prompt": "After the audio ends, extract all occurrences of the word \"lined\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.2, 3.58, "lined"]], "response_mode": "offline", "answer": "[{\"word\":\"lined\",\"start\":3.2,\"end\":3.58}]"} {"utterance_id": "2300-131720-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0028.flac", "task_type": "lexical_counting", "query": "subject", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"subject\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[3.44, 4.04, "one"], [6.78, 7.24, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.44,\"end\":4.04},{\"word\":\"two\",\"start\":6.78,\"end\":7.24}]"} {"utterance_id": "5142-36377-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0012.flac", "task_type": "lexical_detection", "query": "acquaintance", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"acquaintance\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.33, 0.74, "acquaintance"]], "response_mode": "offline", "answer": "[{\"word\":\"acquaintance\",\"start\":0.33,\"end\":0.74}]"} {"utterance_id": "8463-287645-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0005.flac", "task_type": "lexical_detection", "query": "kept", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"kept\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.6, 3.94, "kept"]], "response_mode": "offline", "answer": "[{\"word\":\"kept\",\"start\":3.6,\"end\":3.94}]"} {"utterance_id": "4970-29093-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0012.flac", "task_type": "lexical_counting", "query": "fail", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"fail\" (word, start, end). If none, reply [].", "labels": [[6.35, 6.82, "one"], [7.85, 8.22, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.35,\"end\":6.82},{\"word\":\"two\",\"start\":7.85,\"end\":8.22}]"} {"utterance_id": "8455-210777-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0005.flac", "task_type": "lexical_detection", "query": "elsewhere", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"elsewhere\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.15, 2.75, "elsewhere"]], "response_mode": "offline", "answer": "[{\"word\":\"elsewhere\",\"start\":2.15,\"end\":2.75}]"} {"utterance_id": "7127-75946-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0029.flac", "task_type": "lexical_detection", "query": "quadrille", "prompt": "After the audio ends, extract all occurrences of the word \"quadrille\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.44, 6.09, "quadrille"]], "response_mode": "offline", "answer": "[{\"word\":\"quadrille\",\"start\":5.44,\"end\":6.09}]"} {"utterance_id": "908-157963-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0022.flac", "task_type": "lexical_detection", "query": "forth", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"forth\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 0.81, "forth"]], "response_mode": "offline", "answer": "[{\"word\":\"forth\",\"start\":0.42,\"end\":0.81}]"} {"utterance_id": "121-127105-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0009.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[1.13, 1.62, "twenty"]], "response_mode": "offline", "answer": "[{\"word\":\"twenty\",\"start\":1.13,\"end\":1.62}]"} {"utterance_id": "1188-133604-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0010.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.62, 3.82, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":3.62,\"end\":3.82}]"} {"utterance_id": "7729-102255-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0001.flac", "task_type": "lexical_detection", "query": "march", "prompt": "Once the audio has finished, identify every instance of the word \"march\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.78, 1.18, "march"]], "response_mode": "offline", "answer": "[{\"word\":\"march\",\"start\":0.78,\"end\":1.18}]"} {"utterance_id": "672-122797-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0040.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.96, 2.34, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":1.96,\"end\":2.34}]"} {"utterance_id": "6930-76324-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0018.flac", "task_type": "lexical_detection", "query": "sort", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"sort\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.79, 1.06, "sort"]], "response_mode": "offline", "answer": "[{\"word\":\"sort\",\"start\":0.79,\"end\":1.06}]"} {"utterance_id": "260-123440-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0002.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[5.17, 5.51, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":5.17,\"end\":5.51}]"} {"utterance_id": "260-123286-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0013.flac", "task_type": "lexical_detection", "query": "shadow", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"shadow\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.44, 0.76, "shadow"]], "response_mode": "offline", "answer": "[{\"word\":\"shadow\",\"start\":0.44,\"end\":0.76}]"} {"utterance_id": "1320-122617-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0019.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.39, 0.92, "uncas"]], "response_mode": "offline", "answer": "[{\"word\":\"uncas\",\"start\":0.39,\"end\":0.92}]"} {"utterance_id": "7127-75947-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0014.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[9.23, 9.74, "brain"]], "response_mode": "offline", "answer": "[{\"word\":\"brain\",\"start\":9.23,\"end\":9.74}]"} {"utterance_id": "6829-68771-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0012.flac", "task_type": "lexical_detection", "query": "generation", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"generation\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.43, 3.27, "generation"]], "response_mode": "offline", "answer": "[{\"word\":\"generation\",\"start\":2.43,\"end\":3.27}]"} {"utterance_id": "4446-2275-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0036.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.23, 1.47, "arms"]], "response_mode": "offline", "answer": "[{\"word\":\"arms\",\"start\":1.23,\"end\":1.47}]"} {"utterance_id": "1580-141083-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0028.flac", "task_type": "lexical_detection", "query": "seized", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"seized\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.47, 1.8, "seized"]], "response_mode": "offline", "answer": "[{\"word\":\"seized\",\"start\":1.47,\"end\":1.8}]"} {"utterance_id": "7021-79730-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79730/7021-79730-0006.flac", "task_type": "lexical_detection", "query": "convincing", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"convincing\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.22, 5.97, "convincing"]], "response_mode": "offline", "answer": "[{\"word\":\"convincing\",\"start\":5.22,\"end\":5.97}]"} {"utterance_id": "5142-33396-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0024.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.06, 1.33, "back"]], "response_mode": "offline", "answer": "[{\"word\":\"back\",\"start\":1.06,\"end\":1.33}]"} {"utterance_id": "2961-961-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0002.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Once the audio has finished, list each occurrence of a person's name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[2.51, 3.2, "one"], [8.6, 9.37, "two"], [13.47, 14.44, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.51,\"end\":3.2},{\"word\":\"two\",\"start\":8.6,\"end\":9.37},{\"word\":\"three\",\"start\":13.47,\"end\":14.44}]"} {"utterance_id": "5142-33396-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0053.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.58, 0.93, "five"]], "response_mode": "offline", "answer": "[{\"word\":\"five\",\"start\":0.58,\"end\":0.93}]"} {"utterance_id": "4446-2271-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0003.flac", "task_type": "lexical_detection", "query": "dozen", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"dozen\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.44, 2.68, "dozen"]], "response_mode": "offline", "answer": "[{\"word\":\"dozen\",\"start\":2.44,\"end\":2.68}]"} {"utterance_id": "1188-133604-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0005.flac", "task_type": "lexical_detection", "query": "flower", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"flower\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.23, 2.68, "flower"]], "response_mode": "offline", "answer": "[{\"word\":\"flower\",\"start\":2.23,\"end\":2.68}]"} {"utterance_id": "1580-141084-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0019.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.23, 1.64, "watson"]], "response_mode": "offline", "answer": "[{\"word\":\"watson\",\"start\":1.23,\"end\":1.64}]"} {"utterance_id": "4992-41806-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0002.flac", "task_type": "lexical_detection", "query": "joy", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"joy\" with fields word, start, end (seconds). If none, reply [].", "labels": [[11.62, 12.03, "joy"]], "response_mode": "offline", "answer": "[{\"word\":\"joy\",\"start\":11.62,\"end\":12.03}]"} {"utterance_id": "8224-274381-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0005.flac", "task_type": "lexical_counting", "query": "having", "prompt": "Once the audio has finished, list each occurrence of the word \"having\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[3.95, 4.36, "one"], [19.47, 19.88, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.95,\"end\":4.36},{\"word\":\"two\",\"start\":19.47,\"end\":19.88}]"} {"utterance_id": "8463-287645-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0006.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Process the full audio offline. Find every instance of a profession and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 0.89, "doctor"]], "response_mode": "offline", "answer": "[{\"word\":\"doctor\",\"start\":0.51,\"end\":0.89}]"} {"utterance_id": "1284-1180-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0020.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Process the full audio offline. Find every instance of an animal name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.15, 2.62, "cat"]], "response_mode": "offline", "answer": "[{\"word\":\"cat\",\"start\":2.15,\"end\":2.62}]"} {"utterance_id": "8555-292519-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0007.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.88, 1.44, "heart"], [6.65, 7.1, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":0.88,\"end\":1.44},{\"word\":\"heart\",\"start\":6.65,\"end\":7.1}]"} {"utterance_id": "672-122797-0070", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0070.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the full audio. After it finishes, find every instance of a color name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.65, 1.14, "golden"]], "response_mode": "offline", "answer": "[{\"word\":\"golden\",\"start\":0.65,\"end\":1.14}]"} {"utterance_id": "1221-135767-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0004.flac", "task_type": "semantic_counting", "query": "color", "prompt": "Once the audio has finished, list each occurrence of a color name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[6.49, 7.04, "one"], [14.62, 15.07, "two"], [18.2, 18.74, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.49,\"end\":7.04},{\"word\":\"two\",\"start\":14.62,\"end\":15.07},{\"word\":\"three\",\"start\":18.2,\"end\":18.74}]"} {"utterance_id": "908-31957-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0013.flac", "task_type": "lexical_detection", "query": "angels", "prompt": "When the utterance is complete, report all instances of the word \"angels\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.93, 5.34, "angels"]], "response_mode": "offline", "answer": "[{\"word\":\"angels\",\"start\":4.93,\"end\":5.34}]"} {"utterance_id": "260-123440-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0009.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.39, 1.99, "twenty"]], "response_mode": "offline", "answer": "[{\"word\":\"twenty\",\"start\":1.39,\"end\":1.99}]"} {"utterance_id": "4077-13751-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0015.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.35, 5.83, "thirtieth"]], "response_mode": "offline", "answer": "[{\"word\":\"thirtieth\",\"start\":5.35,\"end\":5.83}]"} {"utterance_id": "1221-135767-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0023.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "Process the full audio offline. For every occurrence of a plant name, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.0, 1.7, "one"], [3.1, 3.85, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.0,\"end\":1.7},{\"word\":\"two\",\"start\":3.1,\"end\":3.85}]"} {"utterance_id": "237-134493-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0006.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen carefully to the entire clip, then report every instance of a city name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.39, 3.73, "town"]], "response_mode": "offline", "answer": "[{\"word\":\"town\",\"start\":3.39,\"end\":3.73}]"} {"utterance_id": "3575-170457-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0030.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of an ordinal number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.9, 1.23, "second"]], "response_mode": "offline", "answer": "[{\"word\":\"second\",\"start\":0.9,\"end\":1.23}]"} {"utterance_id": "121-127105-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0027.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "When the utterance is complete, report all instances of a state or province name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.77, 11.31, "essex"]], "response_mode": "offline", "answer": "[{\"word\":\"essex\",\"start\":10.77,\"end\":11.31}]"} {"utterance_id": "908-157963-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0020.flac", "task_type": "semantic_detection", "query": "color", "prompt": "You will hear a complete utterance. When it ends, list every instance of a color name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.25, 2.74, "golden"]], "response_mode": "offline", "answer": "[{\"word\":\"golden\",\"start\":2.25,\"end\":2.74}]"} {"utterance_id": "5142-33396-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0006.flac", "task_type": "lexical_detection", "query": "follow", "prompt": "After listening to the whole recording, output every instance of the word \"follow\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.21, 3.57, "follow"]], "response_mode": "offline", "answer": "[{\"word\":\"follow\",\"start\":3.21,\"end\":3.57}]"} {"utterance_id": "1188-133604-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0005.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the full audio. Count every occurrence of a person's name. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.98, 4.91, "one"], [7.22, 8.05, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.98,\"end\":4.91},{\"word\":\"two\",\"start\":7.22,\"end\":8.05}]"} {"utterance_id": "2094-142345-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0035.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.72, 3.08, "dinah"]], "response_mode": "offline", "answer": "[{\"word\":\"dinah\",\"start\":2.72,\"end\":3.08}]"} {"utterance_id": "6829-68771-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0007.flac", "task_type": "lexical_detection", "query": "description", "prompt": "Process the full audio offline. Find every instance of the word \"description\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.43, 1.08, "description"]], "response_mode": "offline", "answer": "[{\"word\":\"description\",\"start\":0.43,\"end\":1.08}]"} {"utterance_id": "7176-92135-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0002.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[1.76, 2.33, "dramatist"]], "response_mode": "offline", "answer": "[{\"word\":\"dramatist\",\"start\":1.76,\"end\":2.33}]"} {"utterance_id": "3575-170457-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0042.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.62, 5.08, "surgeon"]], "response_mode": "offline", "answer": "[{\"word\":\"surgeon\",\"start\":4.62,\"end\":5.08}]"} {"utterance_id": "8463-287645-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0012.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.45, 1.94, "jacob's"]], "response_mode": "offline", "answer": "[{\"word\":\"jacob's\",\"start\":1.45,\"end\":1.94}]"} {"utterance_id": "7176-88083-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0024.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "Listen to the entire clip, then report the count sequence for an animal name as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.31, 1.64, "one"], [4.63, 5.06, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.31,\"end\":1.64},{\"word\":\"two\",\"start\":4.63,\"end\":5.06}]"} {"utterance_id": "2830-3979-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3979/2830-3979-0001.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.62, 2.93, "luther"], [11.76, 12.16, "luther"]], "response_mode": "offline", "answer": "[{\"word\":\"luther\",\"start\":2.62,\"end\":2.93},{\"word\":\"luther\",\"start\":11.76,\"end\":12.16}]"} {"utterance_id": "2830-3980-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0018.flac", "task_type": "lexical_detection", "query": "ministry", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"ministry\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.47, 3.01, "ministry"]], "response_mode": "offline", "answer": "[{\"word\":\"ministry\",\"start\":2.47,\"end\":3.01}]"} {"utterance_id": "7176-88083-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0001.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen carefully to the entire clip, then report every instance of an animal name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 1.05, "merganser"]], "response_mode": "offline", "answer": "[{\"word\":\"merganser\",\"start\":0.45,\"end\":1.05}]"} {"utterance_id": "2830-3980-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0007.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "You will hear a complete utterance. When it ends, list every instance of an organization name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.42, 1.86, "government"]], "response_mode": "offline", "answer": "[{\"word\":\"government\",\"start\":1.42,\"end\":1.86}]"} {"utterance_id": "4970-29093-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0006.flac", "task_type": "lexical_detection", "query": "discover", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"discover\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.09, 4.58, "discover"]], "response_mode": "offline", "answer": "[{\"word\":\"discover\",\"start\":4.09,\"end\":4.58}]"} {"utterance_id": "1188-133604-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0015.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Process the full audio offline. Find every instance of a color name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[14.65, 15.1, "golden"]], "response_mode": "offline", "answer": "[{\"word\":\"golden\",\"start\":14.65,\"end\":15.1}]"} {"utterance_id": "6829-68769-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0016.flac", "task_type": "lexical_detection", "query": "tom", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"tom\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.28, 3.66, "tom"]], "response_mode": "offline", "answer": "[{\"word\":\"tom\",\"start\":3.28,\"end\":3.66}]"} {"utterance_id": "672-122797-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0006.flac", "task_type": "lexical_detection", "query": "delightful", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"delightful\" with fields word, start, end (seconds). If none, reply [].", "labels": [[10.94, 11.87, "delightful"]], "response_mode": "offline", "answer": "[{\"word\":\"delightful\",\"start\":10.94,\"end\":11.87}]"} {"utterance_id": "4446-2273-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0027.flac", "task_type": "lexical_detection", "query": "reverie", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"reverie\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.93, 3.54, "reverie"]], "response_mode": "offline", "answer": "[{\"word\":\"reverie\",\"start\":2.93,\"end\":3.54}]"} {"utterance_id": "1580-141084-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0030.flac", "task_type": "lexical_detection", "query": "close", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"close\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.58, 0.84, "close"]], "response_mode": "offline", "answer": "[{\"word\":\"close\",\"start\":0.58,\"end\":0.84}]"} {"utterance_id": "4970-29095-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0026.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Once the audio has finished, identify every instance of a city name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.84, 5.59, "germantown"]], "response_mode": "offline", "answer": "[{\"word\":\"germantown\",\"start\":4.84,\"end\":5.59}]"} {"utterance_id": "1089-134691-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0008.flac", "task_type": "lexical_detection", "query": "strove", "prompt": "After listening to the whole recording, output every instance of the word \"strove\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.94, 7.35, "strove"]], "response_mode": "offline", "answer": "[{\"word\":\"strove\",\"start\":6.94,\"end\":7.35}]"} {"utterance_id": "4077-13754-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0005.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[8.46, 9.03, "brothers"]], "response_mode": "offline", "answer": "[{\"word\":\"brothers\",\"start\":8.46,\"end\":9.03}]"} {"utterance_id": "2094-142345-0060", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0060.flac", "task_type": "lexical_detection", "query": "doubt", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"doubt\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.12, 1.43, "doubt"]], "response_mode": "offline", "answer": "[{\"word\":\"doubt\",\"start\":1.12,\"end\":1.43}]"} {"utterance_id": "8463-294828-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0020.flac", "task_type": "lexical_detection", "query": "replied", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"replied\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.3, 2.75, "replied"]], "response_mode": "offline", "answer": "[{\"word\":\"replied\",\"start\":2.3,\"end\":2.75}]"} {"utterance_id": "1221-135767-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0007.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "After listening to the whole recording, output timed count words for a person's name (one, two, three, ...) as JSON. If none, reply [].", "labels": [[1.51, 1.86, "one"], [6.79, 7.24, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.51,\"end\":1.86},{\"word\":\"two\",\"start\":6.79,\"end\":7.24}]"} {"utterance_id": "1580-141084-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0034.flac", "task_type": "lexical_detection", "query": "listen", "prompt": "Once the audio has finished, identify every instance of the word \"listen\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.89, 2.32, "listen"]], "response_mode": "offline", "answer": "[{\"word\":\"listen\",\"start\":1.89,\"end\":2.32}]"} {"utterance_id": "1320-122617-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0009.flac", "task_type": "lexical_detection", "query": "ignorance", "prompt": "Once the audio has finished, identify every instance of the word \"ignorance\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.45, 3.96, "ignorance"]], "response_mode": "offline", "answer": "[{\"word\":\"ignorance\",\"start\":3.45,\"end\":3.96}]"} {"utterance_id": "2300-131720-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0024.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Process the full audio offline. Find every instance of a country name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.01, 4.46, "country"]], "response_mode": "offline", "answer": "[{\"word\":\"country\",\"start\":4.01,\"end\":4.46}]"} {"utterance_id": "5105-28241-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0016.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Do not answer until the audio ends. Then list each instance of a country name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.28, 4.94, "algeria"]], "response_mode": "offline", "answer": "[{\"word\":\"algeria\",\"start\":4.28,\"end\":4.94}]"} {"utterance_id": "260-123440-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0002.flac", "task_type": "lexical_detection", "query": "rabbit", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"rabbit\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.12, 1.49, "rabbit"]], "response_mode": "offline", "answer": "[{\"word\":\"rabbit\",\"start\":1.12,\"end\":1.49}]"} {"utterance_id": "2961-961-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0012.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.66, 12.02, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":11.66,\"end\":12.02}]"} {"utterance_id": "121-127105-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0034.flac", "task_type": "lexical_counting", "query": "sounded", "prompt": "After listening to the whole recording, output timed count words for the word \"sounded\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[0.53, 0.92, "one"], [1.94, 2.32, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.53,\"end\":0.92},{\"word\":\"two\",\"start\":1.94,\"end\":2.32}]"} {"utterance_id": "61-70968-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0052.flac", "task_type": "lexical_detection", "query": "fellow", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"fellow\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.92, 1.26, "fellow"]], "response_mode": "offline", "answer": "[{\"word\":\"fellow\",\"start\":0.92,\"end\":1.26}]"} {"utterance_id": "2094-142345-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0053.flac", "task_type": "lexical_detection", "query": "send", "prompt": "After listening to the whole recording, output every instance of the word \"send\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.04, 2.31, "send"]], "response_mode": "offline", "answer": "[{\"word\":\"send\",\"start\":2.04,\"end\":2.31}]"} {"utterance_id": "61-70968-0061", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0061.flac", "task_type": "lexical_detection", "query": "worked", "prompt": "After listening to the whole recording, output every instance of the word \"worked\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.38, 4.71, "worked"]], "response_mode": "offline", "answer": "[{\"word\":\"worked\",\"start\":4.38,\"end\":4.71}]"} {"utterance_id": "4992-23283-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0010.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a family relation with fields word, start, end (seconds). If none, reply [].", "labels": [[2.76, 3.21, "husband"]], "response_mode": "offline", "answer": "[{\"word\":\"husband\",\"start\":2.76,\"end\":3.21}]"} {"utterance_id": "237-134493-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0003.flac", "task_type": "lexical_detection", "query": "gilded", "prompt": "Process the full audio offline. Find every instance of the word \"gilded\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.86, 6.33, "gilded"]], "response_mode": "offline", "answer": "[{\"word\":\"gilded\",\"start\":5.86,\"end\":6.33}]"} {"utterance_id": "4446-2273-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0026.flac", "task_type": "lexical_detection", "query": "play", "prompt": "Process the full audio offline. Find every instance of the word \"play\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.4, 1.9, "play"]], "response_mode": "offline", "answer": "[{\"word\":\"play\",\"start\":1.4,\"end\":1.9}]"} {"utterance_id": "1188-133604-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0020.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.22, 3.41, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.22,\"end\":3.41}]"} {"utterance_id": "4970-29095-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0031.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.37, 3.74, "ruth"]], "response_mode": "offline", "answer": "[{\"word\":\"ruth\",\"start\":3.37,\"end\":3.74}]"} {"utterance_id": "8463-294825-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0008.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.81, 3.82, "captain nemo"]], "response_mode": "offline", "answer": "[{\"word\":\"captain nemo\",\"start\":2.81,\"end\":3.82}]"} {"utterance_id": "260-123286-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0027.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a color name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.94, 4.2, "red"]], "response_mode": "offline", "answer": "[{\"word\":\"red\",\"start\":3.94,\"end\":4.2}]"} {"utterance_id": "4992-23283-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0004.flac", "task_type": "lexical_detection", "query": "warranted", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"warranted\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.45, 2.91, "warranted"]], "response_mode": "offline", "answer": "[{\"word\":\"warranted\",\"start\":2.45,\"end\":2.91}]"} {"utterance_id": "4992-41806-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0017.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.72, 2.21, "flowers"], [3.98, 4.76, "corn ears"]], "response_mode": "offline", "answer": "[{\"word\":\"flowers\",\"start\":1.72,\"end\":2.21},{\"word\":\"corn ears\",\"start\":3.98,\"end\":4.76}]"} {"utterance_id": "5683-32879-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0004.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[8.63, 9.05, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":8.63,\"end\":9.05}]"} {"utterance_id": "5683-32865-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0005.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After listening to the whole recording, output every instance of a color name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.32, 9.83, "pale"]], "response_mode": "offline", "answer": "[{\"word\":\"pale\",\"start\":9.32,\"end\":9.83}]"} {"utterance_id": "1580-141084-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0035.flac", "task_type": "lexical_detection", "query": "office", "prompt": "Process the full audio offline. Find every instance of the word \"office\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.79, 2.29, "office"]], "response_mode": "offline", "answer": "[{\"word\":\"office\",\"start\":1.79,\"end\":2.29}]"} {"utterance_id": "8463-294828-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0015.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.9, 2.21, "third"]], "response_mode": "offline", "answer": "[{\"word\":\"third\",\"start\":1.9,\"end\":2.21}]"} {"utterance_id": "8224-274381-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0004.flac", "task_type": "semantic_detection", "query": "city", "prompt": "After listening to the whole recording, output every instance of a city name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.93, 14.36, "perth"]], "response_mode": "offline", "answer": "[{\"word\":\"perth\",\"start\":13.93,\"end\":14.36}]"} {"utterance_id": "8455-210777-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0027.flac", "task_type": "semantic_detection", "query": "country", "prompt": "When the utterance is complete, report all instances of a country name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.3, 3.75, "england"]], "response_mode": "offline", "answer": "[{\"word\":\"england\",\"start\":3.3,\"end\":3.75}]"} {"utterance_id": "61-70968-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0001.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.07, 3.56, "child"]], "response_mode": "offline", "answer": "[{\"word\":\"child\",\"start\":3.07,\"end\":3.56}]"} {"utterance_id": "3729-6852-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0005.flac", "task_type": "lexical_detection", "query": "support", "prompt": "Once the audio has finished, identify every instance of the word \"support\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.07, 5.82, "support"]], "response_mode": "offline", "answer": "[{\"word\":\"support\",\"start\":5.07,\"end\":5.82}]"} {"utterance_id": "1284-1181-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0009.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[1.76, 2.02, "side"]], "response_mode": "offline", "answer": "[{\"word\":\"side\",\"start\":1.76,\"end\":2.02}]"} {"utterance_id": "3570-5695-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0001.flac", "task_type": "semantic_counting", "query": "ordinal", "prompt": "Listen to the entire clip, then report the count sequence for an ordinal number as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.39, 1.75, "one"], [13.91, 14.42, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.39,\"end\":1.75},{\"word\":\"two\",\"start\":13.91,\"end\":14.42}]"} {"utterance_id": "3570-5696-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0001.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a money or currency term with fields word, start, end (seconds). If none, reply [].", "labels": [[11.93, 12.26, "wealth"]], "response_mode": "offline", "answer": "[{\"word\":\"wealth\",\"start\":11.93,\"end\":12.26}]"} {"utterance_id": "7127-75946-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0005.flac", "task_type": "lexical_detection", "query": "inquired", "prompt": "Process the full audio offline. Find every instance of the word \"inquired\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.31, 1.82, "inquired"]], "response_mode": "offline", "answer": "[{\"word\":\"inquired\",\"start\":1.31,\"end\":1.82}]"} {"utterance_id": "7729-102255-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0002.flac", "task_type": "lexical_detection", "query": "states", "prompt": "When the utterance is complete, report all instances of the word \"states\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.03, 4.58, "states"]], "response_mode": "offline", "answer": "[{\"word\":\"states\",\"start\":4.03,\"end\":4.58}]"} {"utterance_id": "260-123288-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0018.flac", "task_type": "lexical_detection", "query": "bears", "prompt": "When the utterance is complete, report all instances of the word \"bears\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.81, 1.27, "bears"]], "response_mode": "offline", "answer": "[{\"word\":\"bears\",\"start\":0.81,\"end\":1.27}]"} {"utterance_id": "7729-102255-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0037.flac", "task_type": "lexical_detection", "query": "property", "prompt": "Process the full audio offline. Find every instance of the word \"property\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.74, 9.28, "property"]], "response_mode": "offline", "answer": "[{\"word\":\"property\",\"start\":8.74,\"end\":9.28}]"} {"utterance_id": "5142-36377-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0017.flac", "task_type": "lexical_detection", "query": "cheerful", "prompt": "Process the full audio offline. Find every instance of the word \"cheerful\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.43, 0.84, "cheerful"]], "response_mode": "offline", "answer": "[{\"word\":\"cheerful\",\"start\":0.43,\"end\":0.84}]"} {"utterance_id": "8463-294828-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0007.flac", "task_type": "lexical_detection", "query": "application", "prompt": "Process the full audio offline. Find every instance of the word \"application\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.88, 9.71, "application"]], "response_mode": "offline", "answer": "[{\"word\":\"application\",\"start\":8.88,\"end\":9.71}]"} {"utterance_id": "260-123288-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0026.flac", "task_type": "lexical_counting", "query": "head", "prompt": "After listening to the whole recording, output timed count words for the word \"head\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[11.65, 12.05, "one"], [14.59, 14.86, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":11.65,\"end\":12.05},{\"word\":\"two\",\"start\":14.59,\"end\":14.86}]"} {"utterance_id": "7176-92135-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0043.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.36, 7.78, "footman"], [11.7, 11.99, "actor"]], "response_mode": "offline", "answer": "[{\"word\":\"footman\",\"start\":7.36,\"end\":7.78},{\"word\":\"actor\",\"start\":11.7,\"end\":11.99}]"} {"utterance_id": "6829-68771-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0001.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "When the utterance is complete, report all instances of a person's name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.82, 1.57, "mister hopkins's"]], "response_mode": "offline", "answer": "[{\"word\":\"mister hopkins's\",\"start\":0.82,\"end\":1.57}]"} {"utterance_id": "2094-142345-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0010.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[4.06, 4.45, "back"]], "response_mode": "offline", "answer": "[{\"word\":\"back\",\"start\":4.06,\"end\":4.45}]"} {"utterance_id": "1995-1837-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0024.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "After listening to the whole recording, output every instance of an animal name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.18, 4.5, "bird"]], "response_mode": "offline", "answer": "[{\"word\":\"bird\",\"start\":4.18,\"end\":4.5}]"} {"utterance_id": "61-70968-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0020.flac", "task_type": "lexical_detection", "query": "shame", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"shame\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.15, 0.65, "shame"]], "response_mode": "offline", "answer": "[{\"word\":\"shame\",\"start\":0.15,\"end\":0.65}]"} {"utterance_id": "6930-76324-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0025.flac", "task_type": "lexical_detection", "query": "peering", "prompt": "When the utterance is complete, report all instances of the word \"peering\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.39, 3.8, "peering"]], "response_mode": "offline", "answer": "[{\"word\":\"peering\",\"start\":3.39,\"end\":3.8}]"} {"utterance_id": "5142-36377-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0025.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a person's name with word/start/end. If none, reply [].", "labels": [[0.39, 0.95, "one"], [2.94, 3.36, "two"], [5.91, 6.34, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.39,\"end\":0.95},{\"word\":\"two\",\"start\":2.94,\"end\":3.36},{\"word\":\"three\",\"start\":5.91,\"end\":6.34}]"} {"utterance_id": "1320-122617-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0005.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.74, 2.39, "sides"]], "response_mode": "offline", "answer": "[{\"word\":\"sides\",\"start\":1.74,\"end\":2.39}]"} {"utterance_id": "908-157963-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0026.flac", "task_type": "lexical_counting", "query": "thee", "prompt": "Listen to the full audio. Count every occurrence of the word \"thee\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.24, 4.57, "one"], [5.6, 5.76, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.24,\"end\":4.57},{\"word\":\"two\",\"start\":5.6,\"end\":5.76}]"} {"utterance_id": "5142-36586-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36586/5142-36586-0003.flac", "task_type": "lexical_detection", "query": "races", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"races\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.55, 3.93, "races"]], "response_mode": "offline", "answer": "[{\"word\":\"races\",\"start\":3.55,\"end\":3.93}]"} {"utterance_id": "8555-292519-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0001.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Process the full audio offline. Find every instance of a color name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[17.58, 17.99, "green"]], "response_mode": "offline", "answer": "[{\"word\":\"green\",\"start\":17.58,\"end\":17.99}]"} {"utterance_id": "1221-135766-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0011.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.28, 1.61, "hester"], [17.45, 17.85, "pearl"]], "response_mode": "offline", "answer": "[{\"word\":\"hester\",\"start\":1.28,\"end\":1.61},{\"word\":\"pearl\",\"start\":17.45,\"end\":17.85}]"} {"utterance_id": "7729-102255-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0038.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.52, 1.26, "atchison"]], "response_mode": "offline", "answer": "[{\"word\":\"atchison\",\"start\":0.52,\"end\":1.26}]"} {"utterance_id": "6829-68769-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0032.flac", "task_type": "lexical_detection", "query": "discovered", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"discovered\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.34, 0.89, "discovered"]], "response_mode": "offline", "answer": "[{\"word\":\"discovered\",\"start\":0.34,\"end\":0.89}]"} {"utterance_id": "1320-122612-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0013.flac", "task_type": "lexical_detection", "query": "portion", "prompt": "After listening to the whole recording, output every instance of the word \"portion\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.63, 6.18, "portion"]], "response_mode": "offline", "answer": "[{\"word\":\"portion\",\"start\":5.63,\"end\":6.18}]"} {"utterance_id": "4970-29093-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0023.flac", "task_type": "semantic_detection", "query": "disease", "prompt": "Listen to the full audio. After it finishes, find every instance of a disease name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.19, 7.72, "fever"]], "response_mode": "offline", "answer": "[{\"word\":\"fever\",\"start\":7.19,\"end\":7.72}]"} {"utterance_id": "2094-142345-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0000.flac", "task_type": "lexical_counting", "query": "red", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"red\" (word, start, end). If none, reply [].", "labels": [[2.38, 2.66, "one"], [12.48, 12.81, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.38,\"end\":2.66},{\"word\":\"two\",\"start\":12.48,\"end\":12.81}]"} {"utterance_id": "2300-131720-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0000.flac", "task_type": "lexical_detection", "query": "palace", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"palace\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.49, 2.99, "palace"]], "response_mode": "offline", "answer": "[{\"word\":\"palace\",\"start\":2.49,\"end\":2.99}]"} {"utterance_id": "5142-36377-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0014.flac", "task_type": "lexical_detection", "query": "america", "prompt": "When the utterance is complete, report all instances of the word \"america\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[20.63, 21.3, "america"]], "response_mode": "offline", "answer": "[{\"word\":\"america\",\"start\":20.63,\"end\":21.3}]"} {"utterance_id": "4077-13754-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0010.flac", "task_type": "semantic_detection", "query": "date", "prompt": "Once the audio has finished, identify every instance of a date. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.56, 1.71, "eighteen sixty two"]], "response_mode": "offline", "answer": "[{\"word\":\"eighteen sixty two\",\"start\":0.56,\"end\":1.71}]"} {"utterance_id": "61-70968-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0009.flac", "task_type": "lexical_detection", "query": "master", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"master\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.58, 2.15, "master"]], "response_mode": "offline", "answer": "[{\"word\":\"master\",\"start\":1.58,\"end\":2.15}]"} {"utterance_id": "2830-3980-0063", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0063.flac", "task_type": "lexical_detection", "query": "comfort", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"comfort\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.91, 3.44, "comfort"]], "response_mode": "offline", "answer": "[{\"word\":\"comfort\",\"start\":2.91,\"end\":3.44}]"} {"utterance_id": "1320-122617-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0025.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.58, 1.09, "uncas"]], "response_mode": "offline", "answer": "[{\"word\":\"uncas\",\"start\":0.58,\"end\":1.09}]"} {"utterance_id": "8455-210777-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0034.flac", "task_type": "lexical_detection", "query": "struggle", "prompt": "When the utterance is complete, report all instances of the word \"struggle\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.46, 6.9, "struggle"]], "response_mode": "offline", "answer": "[{\"word\":\"struggle\",\"start\":6.46,\"end\":6.9}]"} {"utterance_id": "8224-274381-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0003.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.56, 1.01, "king's"]], "response_mode": "offline", "answer": "[{\"word\":\"king's\",\"start\":0.56,\"end\":1.01}]"} {"utterance_id": "61-70970-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0024.flac", "task_type": "lexical_detection", "query": "mission", "prompt": "When the utterance is complete, report all instances of the word \"mission\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.56, 6.99, "mission"]], "response_mode": "offline", "answer": "[{\"word\":\"mission\",\"start\":6.56,\"end\":6.99}]"} {"utterance_id": "5142-36377-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0008.flac", "task_type": "lexical_counting", "query": "defiance", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"defiance\" with word/start/end. If none, reply [].", "labels": [[8.53, 9.16, "one"], [12.89, 13.46, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":8.53,\"end\":9.16},{\"word\":\"two\",\"start\":12.89,\"end\":13.46}]"} {"utterance_id": "1221-135766-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0014.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.51, 0.85, "pearl"]], "response_mode": "offline", "answer": "[{\"word\":\"pearl\",\"start\":0.51,\"end\":0.85}]"} {"utterance_id": "2300-131720-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0018.flac", "task_type": "semantic_counting", "query": "state_or_province", "prompt": "When the utterance is complete, report how a state or province name accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.88, 8.6, "one"], [10.63, 11.12, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.88,\"end\":8.6},{\"word\":\"two\",\"start\":10.63,\"end\":11.12}]"} {"utterance_id": "8463-287645-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/287645/8463-287645-0000.flac", "task_type": "lexical_detection", "query": "concerned", "prompt": "Once the audio has finished, identify every instance of the word \"concerned\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.91, 4.52, "concerned"]], "response_mode": "offline", "answer": "[{\"word\":\"concerned\",\"start\":3.91,\"end\":4.52}]"} {"utterance_id": "6930-81414-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0000.flac", "task_type": "lexical_detection", "query": "existed", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"existed\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.56, 10.17, "existed"]], "response_mode": "offline", "answer": "[{\"word\":\"existed\",\"start\":9.56,\"end\":10.17}]"} {"utterance_id": "4507-16021-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0045.flac", "task_type": "lexical_detection", "query": "punishment", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"punishment\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.41, 4.39, "punishment"]], "response_mode": "offline", "answer": "[{\"word\":\"punishment\",\"start\":3.41,\"end\":4.39}]"} {"utterance_id": "2300-131720-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0004.flac", "task_type": "lexical_counting", "query": "current", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"current\" with word/start/end. If none, reply [].", "labels": [[3.8, 4.18, "one"], [19.0, 19.44, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.8,\"end\":4.18},{\"word\":\"two\",\"start\":19.0,\"end\":19.44}]"} {"utterance_id": "7729-102255-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0015.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.36, 2.01, "generals"], [3.24, 3.87, "governor"]], "response_mode": "offline", "answer": "[{\"word\":\"generals\",\"start\":1.36,\"end\":2.01},{\"word\":\"governor\",\"start\":3.24,\"end\":3.87}]"} {"utterance_id": "7176-88083-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0006.flac", "task_type": "lexical_detection", "query": "fairly", "prompt": "When the utterance is complete, report all instances of the word \"fairly\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.59, 2.97, "fairly"]], "response_mode": "offline", "answer": "[{\"word\":\"fairly\",\"start\":2.59,\"end\":2.97}]"} {"utterance_id": "4507-16021-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0011.flac", "task_type": "lexical_detection", "query": "study", "prompt": "After listening to the whole recording, output every instance of the word \"study\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.61, 4.22, "study"]], "response_mode": "offline", "answer": "[{\"word\":\"study\",\"start\":3.61,\"end\":4.22}]"} {"utterance_id": "121-127105-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0019.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.11, 0.76, "missus griffin"]], "response_mode": "offline", "answer": "[{\"word\":\"missus griffin\",\"start\":0.11,\"end\":0.76}]"} {"utterance_id": "7127-75946-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0006.flac", "task_type": "lexical_detection", "query": "given", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"given\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.62, 0.97, "given"]], "response_mode": "offline", "answer": "[{\"word\":\"given\",\"start\":0.62,\"end\":0.97}]"} {"utterance_id": "4446-2273-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0025.flac", "task_type": "lexical_detection", "query": "take", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"take\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.41, 3.64, "take"]], "response_mode": "offline", "answer": "[{\"word\":\"take\",\"start\":3.41,\"end\":3.64}]"} {"utterance_id": "1221-135767-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0006.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Listen carefully to the entire clip, then report every instance of a color name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.16, 2.47, "red"]], "response_mode": "offline", "answer": "[{\"word\":\"red\",\"start\":2.16,\"end\":2.47}]"} {"utterance_id": "5142-33396-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0035.flac", "task_type": "lexical_detection", "query": "guest", "prompt": "Process the full audio offline. Find every instance of the word \"guest\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.78, 2.11, "guest"]], "response_mode": "offline", "answer": "[{\"word\":\"guest\",\"start\":1.78,\"end\":2.11}]"} {"utterance_id": "1089-134686-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0036.flac", "task_type": "lexical_detection", "query": "xavier", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"xavier\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.47, 3.08, "xavier"]], "response_mode": "offline", "answer": "[{\"word\":\"xavier\",\"start\":2.47,\"end\":3.08}]"} {"utterance_id": "2961-960-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0008.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.47, 9.67, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":9.47,\"end\":9.67}]"} {"utterance_id": "1995-1826-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0023.flac", "task_type": "semantic_counting", "query": "food", "prompt": "After the audio ends, tally every instance of a food name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.29, 0.87, "one"], [5.59, 6.37, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.29,\"end\":0.87},{\"word\":\"two\",\"start\":5.59,\"end\":6.37}]"} {"utterance_id": "5683-32865-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0016.flac", "task_type": "lexical_detection", "query": "brandon", "prompt": "Once the audio has finished, identify every instance of the word \"brandon\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.19, 2.87, "brandon"]], "response_mode": "offline", "answer": "[{\"word\":\"brandon\",\"start\":2.19,\"end\":2.87}]"} {"utterance_id": "61-70968-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0018.flac", "task_type": "lexical_detection", "query": "push", "prompt": "Process the full audio offline. Find every instance of the word \"push\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.14, 1.4, "push"]], "response_mode": "offline", "answer": "[{\"word\":\"push\",\"start\":1.14,\"end\":1.4}]"} {"utterance_id": "4992-23283-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0002.flac", "task_type": "lexical_detection", "query": "minutes", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"minutes\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.76, 2.14, "minutes"]], "response_mode": "offline", "answer": "[{\"word\":\"minutes\",\"start\":1.76,\"end\":2.14}]"} {"utterance_id": "4970-29093-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0017.flac", "task_type": "lexical_detection", "query": "months", "prompt": "When the utterance is complete, report all instances of the word \"months\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.0, 2.49, "months"]], "response_mode": "offline", "answer": "[{\"word\":\"months\",\"start\":2.0,\"end\":2.49}]"} {"utterance_id": "1221-135767-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0011.flac", "task_type": "lexical_detection", "query": "seemingly", "prompt": "When the utterance is complete, report all instances of the word \"seemingly\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.62, 3.16, "seemingly"]], "response_mode": "offline", "answer": "[{\"word\":\"seemingly\",\"start\":2.62,\"end\":3.16}]"} {"utterance_id": "908-31957-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0025.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[15.33, 15.66, "god"]], "response_mode": "offline", "answer": "[{\"word\":\"god\",\"start\":15.33,\"end\":15.66}]"} {"utterance_id": "8463-294828-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0038.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.9, 6.81, "abraham lincoln"]], "response_mode": "offline", "answer": "[{\"word\":\"abraham lincoln\",\"start\":5.9,\"end\":6.81}]"} {"utterance_id": "8230-279154-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0040.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "When the utterance is complete, report all instances of an ordinal number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.35, 8.7, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":8.35,\"end\":8.7}]"} {"utterance_id": "121-123852-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123852/121-123852-0004.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a body part (word, start, end). If none, reply [].", "labels": [[0.53, 1.02, "one"], [7.62, 8.2, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.53,\"end\":1.02},{\"word\":\"two\",\"start\":7.62,\"end\":8.2}]"} {"utterance_id": "8455-210777-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0003.flac", "task_type": "lexical_detection", "query": "mode", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"mode\" with fields word, start, end (seconds). If none, reply [].", "labels": [[5.77, 6.07, "mode"]], "response_mode": "offline", "answer": "[{\"word\":\"mode\",\"start\":5.77,\"end\":6.07}]"} {"utterance_id": "1580-141083-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0006.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[4.31, 4.89, "bannister"]], "response_mode": "offline", "answer": "[{\"word\":\"bannister\",\"start\":4.31,\"end\":4.89}]"} {"utterance_id": "8224-274381-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0004.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "When the utterance is complete, report all instances of an organization name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.58, 4.53, "covenanters"], [9.58, 10.06, "force"], [14.72, 15.14, "army"]], "response_mode": "offline", "answer": "[{\"word\":\"covenanters\",\"start\":3.58,\"end\":4.53},{\"word\":\"force\",\"start\":9.58,\"end\":10.06},{\"word\":\"army\",\"start\":14.72,\"end\":15.14}]"} {"utterance_id": "6829-68771-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0024.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.75, 1.08, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":0.75,\"end\":1.08}]"} {"utterance_id": "237-134493-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0017.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.41, 8.6, "alexandra bergson"]], "response_mode": "offline", "answer": "[{\"word\":\"alexandra bergson\",\"start\":7.41,\"end\":8.6}]"} {"utterance_id": "6829-68771-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0005.flac", "task_type": "lexical_detection", "query": "personally", "prompt": "After listening to the whole recording, output every instance of the word \"personally\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.15, 5.87, "personally"]], "response_mode": "offline", "answer": "[{\"word\":\"personally\",\"start\":5.15,\"end\":5.87}]"} {"utterance_id": "8455-210777-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0029.flac", "task_type": "lexical_detection", "query": "sleep", "prompt": "After the audio ends, extract all occurrences of the word \"sleep\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.42, 2.78, "sleep"]], "response_mode": "offline", "answer": "[{\"word\":\"sleep\",\"start\":2.42,\"end\":2.78}]"} {"utterance_id": "1284-134647-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0005.flac", "task_type": "lexical_detection", "query": "interrupted", "prompt": "Once the audio has finished, identify every instance of the word \"interrupted\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.78, 6.53, "interrupted"]], "response_mode": "offline", "answer": "[{\"word\":\"interrupted\",\"start\":5.78,\"end\":6.53}]"} {"utterance_id": "2830-3980-0042", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0042.flac", "task_type": "lexical_detection", "query": "doctrine", "prompt": "When the utterance is complete, report all instances of the word \"doctrine\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.2, 2.68, "doctrine"]], "response_mode": "offline", "answer": "[{\"word\":\"doctrine\",\"start\":2.2,\"end\":2.68}]"} {"utterance_id": "7729-102255-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0031.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.57, 5.8, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":5.57,\"end\":5.8}]"} {"utterance_id": "5142-36600-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36600/5142-36600-0001.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.08, 1.29, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":1.08,\"end\":1.29}]"} {"utterance_id": "7729-102255-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0036.flac", "task_type": "lexical_detection", "query": "surrender", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"surrender\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.3, 4.01, "surrender"]], "response_mode": "offline", "answer": "[{\"word\":\"surrender\",\"start\":3.3,\"end\":4.01}]"} {"utterance_id": "1320-122617-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0033.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.58, 2.11, "scalp"]], "response_mode": "offline", "answer": "[{\"word\":\"scalp\",\"start\":1.58,\"end\":2.11}]"} {"utterance_id": "8463-294828-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0016.flac", "task_type": "lexical_detection", "query": "entering", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"entering\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.37, 2.92, "entering"]], "response_mode": "offline", "answer": "[{\"word\":\"entering\",\"start\":2.37,\"end\":2.92}]"} {"utterance_id": "3575-170457-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0034.flac", "task_type": "lexical_detection", "query": "mine", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"mine\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.51, 1.85, "mine"]], "response_mode": "offline", "answer": "[{\"word\":\"mine\",\"start\":1.51,\"end\":1.85}]"} {"utterance_id": "2961-961-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0010.flac", "task_type": "lexical_detection", "query": "lands", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"lands\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.74, 4.36, "lands"]], "response_mode": "offline", "answer": "[{\"word\":\"lands\",\"start\":3.74,\"end\":4.36}]"} {"utterance_id": "3575-170457-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0006.flac", "task_type": "lexical_detection", "query": "gathering", "prompt": "After listening to the whole recording, output every instance of the word \"gathering\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.57, 6.08, "gathering"]], "response_mode": "offline", "answer": "[{\"word\":\"gathering\",\"start\":5.57,\"end\":6.08}]"} {"utterance_id": "8224-274384-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0007.flac", "task_type": "lexical_detection", "query": "mercy", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"mercy\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.77, 1.25, "mercy"]], "response_mode": "offline", "answer": "[{\"word\":\"mercy\",\"start\":0.77,\"end\":1.25}]"} {"utterance_id": "8463-294825-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0010.flac", "task_type": "lexical_detection", "query": "classic", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"classic\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.58, 3.06, "classic"]], "response_mode": "offline", "answer": "[{\"word\":\"classic\",\"start\":2.58,\"end\":3.06}]"} {"utterance_id": "3729-6852-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0002.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[7.98, 8.42, "writer"]], "response_mode": "offline", "answer": "[{\"word\":\"writer\",\"start\":7.98,\"end\":8.42}]"} {"utterance_id": "2094-142345-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0026.flac", "task_type": "lexical_detection", "query": "she's", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"she's\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.19, 0.45, "she's"]], "response_mode": "offline", "answer": "[{\"word\":\"she's\",\"start\":0.19,\"end\":0.45}]"} {"utterance_id": "2300-131720-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0024.flac", "task_type": "lexical_detection", "query": "wire", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"wire\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.9, 3.24, "wire"]], "response_mode": "offline", "answer": "[{\"word\":\"wire\",\"start\":2.9,\"end\":3.24}]"} {"utterance_id": "4992-41797-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0013.flac", "task_type": "lexical_detection", "query": "eagerness", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"eagerness\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.55, 3.32, "eagerness"]], "response_mode": "offline", "answer": "[{\"word\":\"eagerness\",\"start\":2.55,\"end\":3.32}]"} {"utterance_id": "672-122797-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0053.flac", "task_type": "lexical_detection", "query": "extremely", "prompt": "After listening to the whole recording, output every instance of the word \"extremely\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.14, 1.9, "extremely"]], "response_mode": "offline", "answer": "[{\"word\":\"extremely\",\"start\":1.14,\"end\":1.9}]"} {"utterance_id": "121-123859-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/123859/121-123859-0000.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[7.73, 8.24, "tongue"]], "response_mode": "offline", "answer": "[{\"word\":\"tongue\",\"start\":7.73,\"end\":8.24}]"} {"utterance_id": "1320-122617-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0041.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.28, 0.69, "uncas"]], "response_mode": "offline", "answer": "[{\"word\":\"uncas\",\"start\":0.28,\"end\":0.69}]"} {"utterance_id": "2300-131720-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0022.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "When the utterance is complete, report all instances of a profession including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.11, 11.7, "inventors"]], "response_mode": "offline", "answer": "[{\"word\":\"inventors\",\"start\":11.11,\"end\":11.7}]"} {"utterance_id": "908-31957-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0019.flac", "task_type": "lexical_detection", "query": "sickness", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"sickness\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.78, 3.42, "sickness"]], "response_mode": "offline", "answer": "[{\"word\":\"sickness\",\"start\":2.78,\"end\":3.42}]"} {"utterance_id": "7729-102255-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0037.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.18, 2.5, "ear"]], "response_mode": "offline", "answer": "[{\"word\":\"ear\",\"start\":2.18,\"end\":2.5}]"} {"utterance_id": "7176-92135-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0032.flac", "task_type": "lexical_detection", "query": "begin", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"begin\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.79, 2.14, "begin"]], "response_mode": "offline", "answer": "[{\"word\":\"begin\",\"start\":1.79,\"end\":2.14}]"} {"utterance_id": "7021-79740-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/79740/7021-79740-0007.flac", "task_type": "lexical_counting", "query": "jane", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"jane\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[1.61, 2.1, "one"], [6.21, 6.65, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.61,\"end\":2.1},{\"word\":\"two\",\"start\":6.21,\"end\":6.65}]"} {"utterance_id": "7176-92135-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0020.flac", "task_type": "lexical_counting", "query": "nine", "prompt": "After the audio ends, tally every instance of the word \"nine\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.72, 1.13, "one"], [3.6, 3.81, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.72,\"end\":1.13},{\"word\":\"two\",\"start\":3.6,\"end\":3.81}]"} {"utterance_id": "3729-6852-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0045.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "After listening to the whole recording, output every instance of an animal name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[16.68, 17.24, "cats"]], "response_mode": "offline", "answer": "[{\"word\":\"cats\",\"start\":16.68,\"end\":17.24}]"} {"utterance_id": "1188-133604-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0010.flac", "task_type": "lexical_detection", "query": "copied", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"copied\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.66, 2.07, "copied"]], "response_mode": "offline", "answer": "[{\"word\":\"copied\",\"start\":1.66,\"end\":2.07}]"} {"utterance_id": "4446-2275-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0033.flac", "task_type": "lexical_detection", "query": "beg", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"beg\" with fields word, start, end (seconds). If none, reply [].", "labels": [[6.35, 6.82, "beg"]], "response_mode": "offline", "answer": "[{\"word\":\"beg\",\"start\":6.35,\"end\":6.82}]"} {"utterance_id": "2830-3980-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0030.flac", "task_type": "lexical_detection", "query": "gospel", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"gospel\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.94, 2.39, "gospel"]], "response_mode": "offline", "answer": "[{\"word\":\"gospel\",\"start\":1.94,\"end\":2.39}]"} {"utterance_id": "2830-3980-0062", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0062.flac", "task_type": "semantic_detection", "query": "country", "prompt": "After listening to the whole recording, output every instance of a country name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.93, 2.5, "kingdom"]], "response_mode": "offline", "answer": "[{\"word\":\"kingdom\",\"start\":1.93,\"end\":2.5}]"} {"utterance_id": "4446-2273-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0010.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "Process the full audio offline. For every occurrence of a plant name, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.63, 1.27, "one"], [7.54, 8.3, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.63,\"end\":1.27},{\"word\":\"two\",\"start\":7.54,\"end\":8.3}]"} {"utterance_id": "1089-134686-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0022.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After listening to the whole recording, output every instance of an ordinal number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.79, 2.21, "first"], [5.95, 6.47, "second"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":1.79,\"end\":2.21},{\"word\":\"second\",\"start\":5.95,\"end\":6.47}]"} {"utterance_id": "4446-2273-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0016.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.42, 0.65, "hair"], [3.02, 3.37, "eyes"], [6.77, 7.18, "nose"]], "response_mode": "offline", "answer": "[{\"word\":\"hair\",\"start\":0.42,\"end\":0.65},{\"word\":\"eyes\",\"start\":3.02,\"end\":3.37},{\"word\":\"nose\",\"start\":6.77,\"end\":7.18}]"} {"utterance_id": "908-157963-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0011.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Process the full audio offline. Find every instance of a plant name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.49, 4.01, "grass"]], "response_mode": "offline", "answer": "[{\"word\":\"grass\",\"start\":3.49,\"end\":4.01}]"} {"utterance_id": "7021-85628-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0012.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.15, 2.46, "head"]], "response_mode": "offline", "answer": "[{\"word\":\"head\",\"start\":2.15,\"end\":2.46}]"} {"utterance_id": "1089-134691-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0019.flac", "task_type": "lexical_detection", "query": "calling", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"calling\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.28, 2.79, "calling"]], "response_mode": "offline", "answer": "[{\"word\":\"calling\",\"start\":2.28,\"end\":2.79}]"} {"utterance_id": "8455-210777-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0014.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.15, 3.61, "wife"]], "response_mode": "offline", "answer": "[{\"word\":\"wife\",\"start\":3.15,\"end\":3.61}]"} {"utterance_id": "4077-13751-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0016.flac", "task_type": "lexical_detection", "query": "resigned", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"resigned\" with fields word, start, end (seconds). If none, reply [].", "labels": [[8.04, 8.56, "resigned"]], "response_mode": "offline", "answer": "[{\"word\":\"resigned\",\"start\":8.04,\"end\":8.56}]"} {"utterance_id": "121-121726-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0010.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[8.16, 8.66, "soldiers"]], "response_mode": "offline", "answer": "[{\"word\":\"soldiers\",\"start\":8.16,\"end\":8.66}]"} {"utterance_id": "1089-134686-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0001.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.76, 2.12, "belly"]], "response_mode": "offline", "answer": "[{\"word\":\"belly\",\"start\":1.76,\"end\":2.12}]"} {"utterance_id": "908-157963-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0029.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Process the full audio offline. Find every instance of a weather term and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.91, 3.4, "wind"]], "response_mode": "offline", "answer": "[{\"word\":\"wind\",\"start\":2.91,\"end\":3.4}]"} {"utterance_id": "61-70970-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0035.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Once the audio has finished, list each occurrence of a person's name as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[0.51, 1.07, "one"], [3.1, 3.49, "two"], [4.54, 4.95, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.51,\"end\":1.07},{\"word\":\"two\",\"start\":3.1,\"end\":3.49},{\"word\":\"three\",\"start\":4.54,\"end\":4.95}]"} {"utterance_id": "672-122797-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0017.flac", "task_type": "lexical_detection", "query": "grander", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"grander\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.97, 2.45, "grander"]], "response_mode": "offline", "answer": "[{\"word\":\"grander\",\"start\":1.97,\"end\":2.45}]"} {"utterance_id": "4992-41806-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0002.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.21, 1.44, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.21,\"end\":1.44}]"} {"utterance_id": "3575-170457-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0026.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.58, 3.9, "second"], [14.42, 14.75, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"second\",\"start\":3.58,\"end\":3.9},{\"word\":\"first\",\"start\":14.42,\"end\":14.75}]"} {"utterance_id": "7176-92135-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0009.flac", "task_type": "lexical_detection", "query": "soliloquy", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"soliloquy\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.3, 4.05, "soliloquy"]], "response_mode": "offline", "answer": "[{\"word\":\"soliloquy\",\"start\":3.3,\"end\":4.05}]"} {"utterance_id": "5142-33396-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0001.flac", "task_type": "lexical_detection", "query": "olaf", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"olaf\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.02, 1.42, "olaf"]], "response_mode": "offline", "answer": "[{\"word\":\"olaf\",\"start\":1.02,\"end\":1.42}]"} {"utterance_id": "2830-3980-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0028.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.82, 2.2, "mouths"]], "response_mode": "offline", "answer": "[{\"word\":\"mouths\",\"start\":1.82,\"end\":2.2}]"} {"utterance_id": "7176-88083-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0022.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for an animal name with word/start/end. If none, reply [].", "labels": [[0.34, 0.72, "one"], [5.43, 5.88, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.34,\"end\":0.72},{\"word\":\"two\",\"start\":5.43,\"end\":5.88}]"} {"utterance_id": "7729-102255-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0007.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "When the utterance is complete, report all instances of an organization name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.98, 5.83, "legislature"]], "response_mode": "offline", "answer": "[{\"word\":\"legislature\",\"start\":4.98,\"end\":5.83}]"} {"utterance_id": "3570-5694-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0016.flac", "task_type": "lexical_detection", "query": "attached", "prompt": "After the audio ends, extract all occurrences of the word \"attached\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.12, 5.74, "attached"]], "response_mode": "offline", "answer": "[{\"word\":\"attached\",\"start\":5.12,\"end\":5.74}]"} {"utterance_id": "4507-16021-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0030.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "After listening to the whole recording, output every instance of a state or province name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.31, 4.84, "province"]], "response_mode": "offline", "answer": "[{\"word\":\"province\",\"start\":4.31,\"end\":4.84}]"} {"utterance_id": "61-70968-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0002.flac", "task_type": "lexical_detection", "query": "fortune", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"fortune\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.8, 1.35, "fortune"]], "response_mode": "offline", "answer": "[{\"word\":\"fortune\",\"start\":0.8,\"end\":1.35}]"} {"utterance_id": "5639-40744-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0041.flac", "task_type": "lexical_detection", "query": "assure", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"assure\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.84, 4.2, "assure"]], "response_mode": "offline", "answer": "[{\"word\":\"assure\",\"start\":3.84,\"end\":4.2}]"} {"utterance_id": "5105-28241-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0005.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.72, 0.92, "few"]], "response_mode": "offline", "answer": "[{\"word\":\"few\",\"start\":0.72,\"end\":0.92}]"} {"utterance_id": "4507-16021-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0034.flac", "task_type": "lexical_detection", "query": "orders", "prompt": "After the audio ends, extract all occurrences of the word \"orders\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.1, 2.69, "orders"]], "response_mode": "offline", "answer": "[{\"word\":\"orders\",\"start\":2.1,\"end\":2.69}]"} {"utterance_id": "7176-92135-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0008.flac", "task_type": "lexical_detection", "query": "lend", "prompt": "After listening to the whole recording, output every instance of the word \"lend\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.15, 0.44, "lend"]], "response_mode": "offline", "answer": "[{\"word\":\"lend\",\"start\":0.15,\"end\":0.44}]"} {"utterance_id": "4446-2275-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0034.flac", "task_type": "lexical_detection", "query": "keep", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"keep\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.2, 0.45, "keep"]], "response_mode": "offline", "answer": "[{\"word\":\"keep\",\"start\":0.2,\"end\":0.45}]"} {"utterance_id": "4992-41797-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0000.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[1.9, 2.24, "four"]], "response_mode": "offline", "answer": "[{\"word\":\"four\",\"start\":1.9,\"end\":2.24}]"} {"utterance_id": "61-70970-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0006.flac", "task_type": "lexical_detection", "query": "sir", "prompt": "Process the full audio offline. Find every instance of the word \"sir\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.05, 1.39, "sir"]], "response_mode": "offline", "answer": "[{\"word\":\"sir\",\"start\":1.05,\"end\":1.39}]"} {"utterance_id": "8455-210777-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0031.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Once the audio has finished, identify every instance of a country name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.29, 5.73, "england"]], "response_mode": "offline", "answer": "[{\"word\":\"england\",\"start\":5.29,\"end\":5.73}]"} {"utterance_id": "7127-75947-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0023.flac", "task_type": "lexical_detection", "query": "loudly", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"loudly\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.37, 4.87, "loudly"]], "response_mode": "offline", "answer": "[{\"word\":\"loudly\",\"start\":4.37,\"end\":4.87}]"} {"utterance_id": "2300-131720-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0032.flac", "task_type": "lexical_counting", "query": "meter", "prompt": "Once the audio has finished, list each occurrence of the word \"meter\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[1.84, 2.13, "one"], [5.21, 5.53, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.84,\"end\":2.13},{\"word\":\"two\",\"start\":5.21,\"end\":5.53}]"} {"utterance_id": "4970-29095-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0003.flac", "task_type": "lexical_detection", "query": "loves", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"loves\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[6.44, 6.74, "loves"]], "response_mode": "offline", "answer": "[{\"word\":\"loves\",\"start\":6.44,\"end\":6.74}]"} {"utterance_id": "6930-81414-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0017.flac", "task_type": "lexical_detection", "query": "evening", "prompt": "After the audio ends, extract all occurrences of the word \"evening\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.71, 2.23, "evening"]], "response_mode": "offline", "answer": "[{\"word\":\"evening\",\"start\":1.71,\"end\":2.23}]"} {"utterance_id": "5639-40744-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0003.flac", "task_type": "lexical_counting", "query": "design", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"design\" (word, start, end). If none, reply [].", "labels": [[17.46, 18.0, "one"], [23.74, 24.23, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":17.46,\"end\":18.0},{\"word\":\"two\",\"start\":23.74,\"end\":24.23}]"} {"utterance_id": "7176-92135-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0008.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.09, 1.33, "ten"]], "response_mode": "offline", "answer": "[{\"word\":\"ten\",\"start\":1.09,\"end\":1.33}]"} {"utterance_id": "237-134500-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0026.flac", "task_type": "lexical_detection", "query": "surely", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"surely\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.24, 0.6, "surely"]], "response_mode": "offline", "answer": "[{\"word\":\"surely\",\"start\":0.24,\"end\":0.6}]"} {"utterance_id": "237-126133-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0019.flac", "task_type": "lexical_detection", "query": "ejaculated", "prompt": "After listening to the whole recording, output every instance of the word \"ejaculated\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.41, 2.18, "ejaculated"]], "response_mode": "offline", "answer": "[{\"word\":\"ejaculated\",\"start\":1.41,\"end\":2.18}]"} {"utterance_id": "1580-141083-0052", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0052.flac", "task_type": "lexical_detection", "query": "sir", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"sir\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.58, 3.25, "sir"]], "response_mode": "offline", "answer": "[{\"word\":\"sir\",\"start\":2.58,\"end\":3.25}]"} {"utterance_id": "5683-32879-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0016.flac", "task_type": "lexical_detection", "query": "cousin's", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"cousin's\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.7, 9.19, "cousin's"]], "response_mode": "offline", "answer": "[{\"word\":\"cousin's\",\"start\":8.7,\"end\":9.19}]"} {"utterance_id": "672-122797-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0043.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After the audio ends, extract all occurrences of a plant name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.97, 3.38, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":2.97,\"end\":3.38}]"} {"utterance_id": "4992-41797-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0019.flac", "task_type": "semantic_counting", "query": "color", "prompt": "Listen to the entire clip, then report the count sequence for a color name as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.66, 2.01, "one"], [3.72, 4.11, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.66,\"end\":2.01},{\"word\":\"two\",\"start\":3.72,\"end\":4.11}]"} {"utterance_id": "8224-274381-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274381/8224-274381-0004.flac", "task_type": "lexical_counting", "query": "men", "prompt": "Once the audio has finished, list each occurrence of the word \"men\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[1.08, 1.35, "one"], [16.16, 16.58, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.08,\"end\":1.35},{\"word\":\"two\",\"start\":16.16,\"end\":16.58}]"} {"utterance_id": "1580-141084-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0007.flac", "task_type": "lexical_detection", "query": "morrow", "prompt": "Process the full audio offline. Find every instance of the word \"morrow\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.37, 0.73, "morrow"]], "response_mode": "offline", "answer": "[{\"word\":\"morrow\",\"start\":0.37,\"end\":0.73}]"} {"utterance_id": "5683-32865-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0011.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.37, 6.08, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":5.37,\"end\":6.08}]"} {"utterance_id": "7176-92135-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0040.flac", "task_type": "lexical_detection", "query": "cup", "prompt": "When the utterance is complete, report all instances of the word \"cup\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.53, 5.81, "cup"]], "response_mode": "offline", "answer": "[{\"word\":\"cup\",\"start\":5.53,\"end\":5.81}]"} {"utterance_id": "8463-294828-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0006.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.61, 7.0, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":6.61,\"end\":7.0}]"} {"utterance_id": "237-134500-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0018.flac", "task_type": "lexical_detection", "query": "colored", "prompt": "After the audio ends, extract all occurrences of the word \"colored\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.39, 6.72, "colored"]], "response_mode": "offline", "answer": "[{\"word\":\"colored\",\"start\":6.39,\"end\":6.72}]"} {"utterance_id": "7729-102255-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0034.flac", "task_type": "lexical_detection", "query": "sorrow", "prompt": "After listening to the whole recording, output every instance of the word \"sorrow\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.55, 0.97, "sorrow"]], "response_mode": "offline", "answer": "[{\"word\":\"sorrow\",\"start\":0.55,\"end\":0.97}]"} {"utterance_id": "8463-294828-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0015.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a person's name with word/start/end. If none, reply [].", "labels": [[0.53, 0.93, "one"], [3.38, 3.95, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.53,\"end\":0.93},{\"word\":\"two\",\"start\":3.38,\"end\":3.95}]"} {"utterance_id": "7127-75946-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0015.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an organization name with fields word, start, end (seconds). If none, reply [].", "labels": [[5.68, 6.18, "court"]], "response_mode": "offline", "answer": "[{\"word\":\"court\",\"start\":5.68,\"end\":6.18}]"} {"utterance_id": "260-123288-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0008.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "Process the full audio offline. Find every instance of a weather term and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.18, 1.71, "storm"]], "response_mode": "offline", "answer": "[{\"word\":\"storm\",\"start\":1.18,\"end\":1.71}]"} {"utterance_id": "260-123286-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0026.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "After the audio ends, tally every instance of a number. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.13, 0.4, "one"], [4.68, 4.95, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.13,\"end\":0.4},{\"word\":\"two\",\"start\":4.68,\"end\":4.95}]"} {"utterance_id": "5142-33396-0047", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0047.flac", "task_type": "lexical_detection", "query": "table", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"table\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.16, 1.46, "table"]], "response_mode": "offline", "answer": "[{\"word\":\"table\",\"start\":1.16,\"end\":1.46}]"} {"utterance_id": "7127-75947-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0014.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.02, 3.42, "thousand"]], "response_mode": "offline", "answer": "[{\"word\":\"thousand\",\"start\":3.02,\"end\":3.42}]"} {"utterance_id": "2961-961-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.51, 3.2, "timaeus"], [8.6, 9.37, "critias"], [13.47, 14.44, "hermocrates"]], "response_mode": "offline", "answer": "[{\"word\":\"timaeus\",\"start\":2.51,\"end\":3.2},{\"word\":\"critias\",\"start\":8.6,\"end\":9.37},{\"word\":\"hermocrates\",\"start\":13.47,\"end\":14.44}]"} {"utterance_id": "1580-141083-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0036.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.26, 0.58, "holmes"]], "response_mode": "offline", "answer": "[{\"word\":\"holmes\",\"start\":0.26,\"end\":0.58}]"} {"utterance_id": "1580-141083-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0013.flac", "task_type": "lexical_detection", "query": "desire", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"desire\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.23, 1.58, "desire"]], "response_mode": "offline", "answer": "[{\"word\":\"desire\",\"start\":1.23,\"end\":1.58}]"} {"utterance_id": "7021-85628-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0019.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.45, 1.94, "anders"]], "response_mode": "offline", "answer": "[{\"word\":\"anders\",\"start\":1.45,\"end\":1.94}]"} {"utterance_id": "2830-3980-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0035.flac", "task_type": "lexical_detection", "query": "men's", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"men's\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.1, 2.35, "men's"]], "response_mode": "offline", "answer": "[{\"word\":\"men's\",\"start\":2.1,\"end\":2.35}]"} {"utterance_id": "1284-134647-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/134647/1284-134647-0007.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a profession (word, start, end). If none, reply [].", "labels": [[10.87, 11.29, "one"], [13.1, 13.69, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":10.87,\"end\":11.29},{\"word\":\"two\",\"start\":13.1,\"end\":13.69}]"} {"utterance_id": "4446-2273-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0021.flac", "task_type": "lexical_detection", "query": "youth", "prompt": "When the utterance is complete, report all instances of the word \"youth\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.5, 5.01, "youth"]], "response_mode": "offline", "answer": "[{\"word\":\"youth\",\"start\":4.5,\"end\":5.01}]"} {"utterance_id": "1284-1181-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0003.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Once the audio has finished, list each occurrence of a body part as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[0.36, 0.67, "one"], [2.8, 3.11, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.36,\"end\":0.67},{\"word\":\"two\",\"start\":2.8,\"end\":3.11}]"} {"utterance_id": "121-127105-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0016.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the full audio. After it finishes, find every instance of an ordinal number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.1, 1.44, "second"]], "response_mode": "offline", "answer": "[{\"word\":\"second\",\"start\":1.1,\"end\":1.44}]"} {"utterance_id": "7127-75946-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0028.flac", "task_type": "lexical_detection", "query": "degrees", "prompt": "Once the audio has finished, identify every instance of the word \"degrees\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.67, 1.22, "degrees"]], "response_mode": "offline", "answer": "[{\"word\":\"degrees\",\"start\":0.67,\"end\":1.22}]"} {"utterance_id": "3729-6852-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0005.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.49, 1.08, "silvia"], [12.18, 13.03, "marivaux"]], "response_mode": "offline", "answer": "[{\"word\":\"silvia\",\"start\":0.49,\"end\":1.08},{\"word\":\"marivaux\",\"start\":12.18,\"end\":13.03}]"} {"utterance_id": "1221-135767-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0020.flac", "task_type": "lexical_detection", "query": "hidden", "prompt": "Once the audio has finished, identify every instance of the word \"hidden\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.06, 2.44, "hidden"]], "response_mode": "offline", "answer": "[{\"word\":\"hidden\",\"start\":2.06,\"end\":2.44}]"} {"utterance_id": "5105-28240-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28240/5105-28240-0013.flac", "task_type": "lexical_detection", "query": "yourself", "prompt": "After listening to the whole recording, output every instance of the word \"yourself\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.76, 2.46, "yourself"]], "response_mode": "offline", "answer": "[{\"word\":\"yourself\",\"start\":1.76,\"end\":2.46}]"} {"utterance_id": "5683-32865-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0012.flac", "task_type": "lexical_detection", "query": "perceive", "prompt": "Process the full audio offline. Find every instance of the word \"perceive\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.54, 13.98, "perceive"]], "response_mode": "offline", "answer": "[{\"word\":\"perceive\",\"start\":13.54,\"end\":13.98}]"} {"utterance_id": "7176-92135-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0040.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After listening to the whole recording, output every instance of a color name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.22, 6.47, "brown"]], "response_mode": "offline", "answer": "[{\"word\":\"brown\",\"start\":6.22,\"end\":6.47}]"} {"utterance_id": "4992-41806-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0005.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Process the full audio offline. Find every instance of an ordinal number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.54, 0.91, "next"]], "response_mode": "offline", "answer": "[{\"word\":\"next\",\"start\":0.54,\"end\":0.91}]"} {"utterance_id": "1221-135767-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0024.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a plant name: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[1.37, 2.1, "one"], [3.5, 3.92, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.37,\"end\":2.1},{\"word\":\"two\",\"start\":3.5,\"end\":3.92}]"} {"utterance_id": "6829-68769-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0006.flac", "task_type": "lexical_detection", "query": "withdrawn", "prompt": "Once the audio has finished, identify every instance of the word \"withdrawn\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.59, 2.23, "withdrawn"]], "response_mode": "offline", "answer": "[{\"word\":\"withdrawn\",\"start\":1.59,\"end\":2.23}]"} {"utterance_id": "908-31957-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0005.flac", "task_type": "lexical_detection", "query": "alas", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"alas\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.27, 0.98, "alas"]], "response_mode": "offline", "answer": "[{\"word\":\"alas\",\"start\":0.27,\"end\":0.98}]"} {"utterance_id": "6829-68771-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0035.flac", "task_type": "lexical_detection", "query": "away", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"away\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.78, 2.99, "away"]], "response_mode": "offline", "answer": "[{\"word\":\"away\",\"start\":2.78,\"end\":2.99}]"} {"utterance_id": "2830-3980-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.04, 6.38, "christ"]], "response_mode": "offline", "answer": "[{\"word\":\"christ\",\"start\":6.04,\"end\":6.38}]"} {"utterance_id": "1089-134691-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0005.flac", "task_type": "lexical_detection", "query": "everlasting", "prompt": "Once the audio has finished, identify every instance of the word \"everlasting\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.83, 4.57, "everlasting"]], "response_mode": "offline", "answer": "[{\"word\":\"everlasting\",\"start\":3.83,\"end\":4.57}]"} {"utterance_id": "1284-1181-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0004.flac", "task_type": "lexical_detection", "query": "purposes", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"purposes\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.32, 4.92, "purposes"]], "response_mode": "offline", "answer": "[{\"word\":\"purposes\",\"start\":4.32,\"end\":4.92}]"} {"utterance_id": "8455-210777-0070", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0070.flac", "task_type": "lexical_detection", "query": "adopted", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"adopted\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.0, 1.48, "adopted"]], "response_mode": "offline", "answer": "[{\"word\":\"adopted\",\"start\":1.0,\"end\":1.48}]"} {"utterance_id": "2830-3980-0057", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0057.flac", "task_type": "lexical_detection", "query": "created", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"created\" with fields word, start, end (seconds). If none, reply [].", "labels": [[6.46, 6.81, "created"]], "response_mode": "offline", "answer": "[{\"word\":\"created\",\"start\":6.46,\"end\":6.81}]"} {"utterance_id": "5105-28233-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28233/5105-28233-0005.flac", "task_type": "lexical_detection", "query": "wander", "prompt": "When the utterance is complete, report all instances of the word \"wander\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.17, 1.63, "wander"]], "response_mode": "offline", "answer": "[{\"word\":\"wander\",\"start\":1.17,\"end\":1.63}]"} {"utterance_id": "1221-135766-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0010.flac", "task_type": "lexical_detection", "query": "inexplicable", "prompt": "Once the audio has finished, identify every instance of the word \"inexplicable\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.44, 3.39, "inexplicable"]], "response_mode": "offline", "answer": "[{\"word\":\"inexplicable\",\"start\":2.44,\"end\":3.39}]"} {"utterance_id": "3575-170457-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0032.flac", "task_type": "lexical_detection", "query": "getting", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"getting\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.94, 1.2, "getting"]], "response_mode": "offline", "answer": "[{\"word\":\"getting\",\"start\":0.94,\"end\":1.2}]"} {"utterance_id": "8230-279154-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0035.flac", "task_type": "lexical_detection", "query": "familiarity", "prompt": "After the audio ends, extract all occurrences of the word \"familiarity\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.27, 6.21, "familiarity"]], "response_mode": "offline", "answer": "[{\"word\":\"familiarity\",\"start\":5.27,\"end\":6.21}]"} {"utterance_id": "7127-75946-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0014.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.9, "queens"]], "response_mode": "offline", "answer": "[{\"word\":\"queens\",\"start\":0.53,\"end\":0.9}]"} {"utterance_id": "6829-68771-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0030.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.21, 3.52, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.21,\"end\":3.52}]"} {"utterance_id": "5142-33396-0063", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0063.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Once the audio has finished, identify every instance of a family relation. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.78, 2.31, "baby"]], "response_mode": "offline", "answer": "[{\"word\":\"baby\",\"start\":1.78,\"end\":2.31}]"} {"utterance_id": "7729-102255-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0037.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen carefully to the entire clip, then report every instance of a profession as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.5, 5.39, "proprietors"]], "response_mode": "offline", "answer": "[{\"word\":\"proprietors\",\"start\":4.5,\"end\":5.39}]"} {"utterance_id": "1188-133604-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0041.flac", "task_type": "lexical_detection", "query": "stones", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"stones\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.57, 8.99, "stones"]], "response_mode": "offline", "answer": "[{\"word\":\"stones\",\"start\":8.57,\"end\":8.99}]"} {"utterance_id": "1580-141083-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0022.flac", "task_type": "lexical_detection", "query": "signs", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"signs\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.54, 1.97, "signs"]], "response_mode": "offline", "answer": "[{\"word\":\"signs\",\"start\":1.54,\"end\":1.97}]"} {"utterance_id": "260-123286-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0031.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Do not answer until the audio ends. Then list each instance of an animal name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.09, 2.19, "ichthyosaurus"]], "response_mode": "offline", "answer": "[{\"word\":\"ichthyosaurus\",\"start\":1.09,\"end\":2.19}]"} {"utterance_id": "6829-68771-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0015.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.1, 2.39, "martha"]], "response_mode": "offline", "answer": "[{\"word\":\"martha\",\"start\":2.1,\"end\":2.39}]"} {"utterance_id": "1580-141084-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0015.flac", "task_type": "lexical_detection", "query": "impressed", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"impressed\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.65, 1.07, "impressed"]], "response_mode": "offline", "answer": "[{\"word\":\"impressed\",\"start\":0.65,\"end\":1.07}]"} {"utterance_id": "260-123440-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0013.flac", "task_type": "lexical_detection", "query": "alone", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"alone\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.3, 2.66, "alone"]], "response_mode": "offline", "answer": "[{\"word\":\"alone\",\"start\":2.3,\"end\":2.66}]"} {"utterance_id": "1188-133604-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0039.flac", "task_type": "lexical_detection", "query": "lines", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"lines\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.61, 5.04, "lines"]], "response_mode": "offline", "answer": "[{\"word\":\"lines\",\"start\":4.61,\"end\":5.04}]"} {"utterance_id": "237-134500-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0018.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.37, 2.04, "branches"], [4.36, 4.77, "fruit"], [6.72, 7.24, "berries"]], "response_mode": "offline", "answer": "[{\"word\":\"branches\",\"start\":1.37,\"end\":2.04},{\"word\":\"fruit\",\"start\":4.36,\"end\":4.77},{\"word\":\"berries\",\"start\":6.72,\"end\":7.24}]"} {"utterance_id": "1995-1837-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0023.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[11.96, 12.46, "fingers"], [15.6, 16.02, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"fingers\",\"start\":11.96,\"end\":12.46},{\"word\":\"eyes\",\"start\":15.6,\"end\":16.02}]"} {"utterance_id": "121-127105-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0028.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.12, 3.8, "relations"]], "response_mode": "offline", "answer": "[{\"word\":\"relations\",\"start\":3.12,\"end\":3.8}]"} {"utterance_id": "1284-1180-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0020.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Do not answer until the audio ends. Then list each instance of an ordinal number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.27, 0.67, "first"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":0.27,\"end\":0.67}]"} {"utterance_id": "7127-75946-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0002.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After listening to the whole recording, output every instance of a profession with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.94, 2.41, "king"], [10.24, 10.48, "king"], [12.05, 12.78, "courtiers"]], "response_mode": "offline", "answer": "[{\"word\":\"king\",\"start\":1.94,\"end\":2.41},{\"word\":\"king\",\"start\":10.24,\"end\":10.48},{\"word\":\"courtiers\",\"start\":12.05,\"end\":12.78}]"} {"utterance_id": "1580-141083-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0009.flac", "task_type": "lexical_detection", "query": "passing", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"passing\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.76, 2.25, "passing"]], "response_mode": "offline", "answer": "[{\"word\":\"passing\",\"start\":1.76,\"end\":2.25}]"} {"utterance_id": "8555-284447-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0023.flac", "task_type": "lexical_detection", "query": "shiver", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"shiver\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.72, 1.05, "shiver"], [5.17, 5.54, "shiver"]], "response_mode": "offline", "answer": "[{\"word\":\"shiver\",\"start\":0.72,\"end\":1.05},{\"word\":\"shiver\",\"start\":5.17,\"end\":5.54}]"} {"utterance_id": "4507-16021-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0023.flac", "task_type": "lexical_counting", "query": "manufacturer", "prompt": "Process the full audio offline. For every occurrence of the word \"manufacturer\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.91, 2.05, "one"], [9.5, 10.61, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.91,\"end\":2.05},{\"word\":\"two\",\"start\":9.5,\"end\":10.61}]"} {"utterance_id": "5639-40744-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0036.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.38, 0.68, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":0.38,\"end\":0.68}]"} {"utterance_id": "4507-16021-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0026.flac", "task_type": "lexical_detection", "query": "civilization", "prompt": "After listening to the whole recording, output every instance of the word \"civilization\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[18.43, 19.59, "civilization"], [29.13, 30.43, "civilization"]], "response_mode": "offline", "answer": "[{\"word\":\"civilization\",\"start\":18.43,\"end\":19.59},{\"word\":\"civilization\",\"start\":29.13,\"end\":30.43}]"} {"utterance_id": "4077-13751-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0014.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Do not answer until the audio ends. Then list each instance of a state or province name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.67, 6.36, "clay county"], [8.76, 9.75, "caldwell county"]], "response_mode": "offline", "answer": "[{\"word\":\"clay county\",\"start\":5.67,\"end\":6.36},{\"word\":\"caldwell county\",\"start\":8.76,\"end\":9.75}]"} {"utterance_id": "2300-131720-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0021.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.12, 8.51, "edison"]], "response_mode": "offline", "answer": "[{\"word\":\"edison\",\"start\":8.12,\"end\":8.51}]"} {"utterance_id": "7176-92135-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0014.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the full audio. After it finishes, find every instance of an ordinal number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.1, 1.61, "first"], [5.41, 6.07, "secondly"]], "response_mode": "offline", "answer": "[{\"word\":\"first\",\"start\":1.1,\"end\":1.61},{\"word\":\"secondly\",\"start\":5.41,\"end\":6.07}]"} {"utterance_id": "6930-76324-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0003.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.61, 1.86, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":1.61,\"end\":1.86}]"} {"utterance_id": "7127-75946-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0005.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.82, 2.31, "louis"]], "response_mode": "offline", "answer": "[{\"word\":\"louis\",\"start\":1.82,\"end\":2.31}]"} {"utterance_id": "260-123286-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0026.flac", "task_type": "lexical_detection", "query": "monsters", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"monsters\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.4, 0.94, "monsters"]], "response_mode": "offline", "answer": "[{\"word\":\"monsters\",\"start\":0.4,\"end\":0.94}]"} {"utterance_id": "5142-33396-0059", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0059.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen carefully to the entire clip, then report every instance of a body part as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.47, 1.86, "fingers"]], "response_mode": "offline", "answer": "[{\"word\":\"fingers\",\"start\":1.47,\"end\":1.86}]"} {"utterance_id": "7176-88083-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0003.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.06, 7.46, "claws"]], "response_mode": "offline", "answer": "[{\"word\":\"claws\",\"start\":7.06,\"end\":7.46}]"} {"utterance_id": "1284-1180-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0007.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Do not answer until the audio ends. Then list each instance of a profession as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.41, 3.03, "magician"]], "response_mode": "offline", "answer": "[{\"word\":\"magician\",\"start\":2.41,\"end\":3.03}]"} {"utterance_id": "4507-16021-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0018.flac", "task_type": "lexical_detection", "query": "properly", "prompt": "After listening to the whole recording, output every instance of the word \"properly\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.97, 2.45, "properly"]], "response_mode": "offline", "answer": "[{\"word\":\"properly\",\"start\":1.97,\"end\":2.45}]"} {"utterance_id": "3570-5694-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0008.flac", "task_type": "lexical_counting", "query": "consumption", "prompt": "When the utterance is complete, report how the word \"consumption\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.59, 1.14, "one"], [3.41, 4.02, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.59,\"end\":1.14},{\"word\":\"two\",\"start\":3.41,\"end\":4.02}]"} {"utterance_id": "4992-41806-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0014.flac", "task_type": "lexical_detection", "query": "mend", "prompt": "When the utterance is complete, report all instances of the word \"mend\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.9, 4.26, "mend"]], "response_mode": "offline", "answer": "[{\"word\":\"mend\",\"start\":3.9,\"end\":4.26}]"} {"utterance_id": "1188-133604-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0005.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen carefully to the entire clip, then report every instance of a number as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.78, 5.94, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.78,\"end\":5.94}]"} {"utterance_id": "260-123286-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0009.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "You will hear a complete utterance. When it ends, list every instance of a profession you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.54, 3.13, "professor"]], "response_mode": "offline", "answer": "[{\"word\":\"professor\",\"start\":2.54,\"end\":3.13}]"} {"utterance_id": "5142-33396-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0018.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a money or currency term with fields word, start, end (seconds). If none, reply [].", "labels": [[3.25, 3.71, "gold"]], "response_mode": "offline", "answer": "[{\"word\":\"gold\",\"start\":3.25,\"end\":3.71}]"} {"utterance_id": "6930-81414-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/81414/6930-81414-0026.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.36, 1.84, "nerves"]], "response_mode": "offline", "answer": "[{\"word\":\"nerves\",\"start\":1.36,\"end\":1.84}]"} {"utterance_id": "1089-134686-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0002.flac", "task_type": "lexical_detection", "query": "lamps", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"lamps\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.16, 2.59, "lamps"]], "response_mode": "offline", "answer": "[{\"word\":\"lamps\",\"start\":2.16,\"end\":2.59}]"} {"utterance_id": "672-122797-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0023.flac", "task_type": "lexical_detection", "query": "birds", "prompt": "After listening to the whole recording, output every instance of the word \"birds\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.57, 9.19, "birds"]], "response_mode": "offline", "answer": "[{\"word\":\"birds\",\"start\":8.57,\"end\":9.19}]"} {"utterance_id": "5105-28233-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28233/5105-28233-0010.flac", "task_type": "lexical_counting", "query": "made", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"made\" with word/start/end. If none, reply [].", "labels": [[1.56, 1.86, "one"], [6.92, 7.17, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.56,\"end\":1.86},{\"word\":\"two\",\"start\":6.92,\"end\":7.17}]"} {"utterance_id": "7176-88083-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0011.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a body part with word/start/end. If none, reply [].", "labels": [[3.58, 3.96, "one"], [6.56, 6.91, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.58,\"end\":3.96},{\"word\":\"two\",\"start\":6.56,\"end\":6.91}]"} {"utterance_id": "3575-170457-0055", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0055.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.06, 1.3, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":1.06,\"end\":1.3}]"} {"utterance_id": "4446-2271-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0017.flac", "task_type": "lexical_detection", "query": "studying", "prompt": "After the audio ends, extract all occurrences of the word \"studying\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.17, 5.55, "studying"]], "response_mode": "offline", "answer": "[{\"word\":\"studying\",\"start\":5.17,\"end\":5.55}]"} {"utterance_id": "672-122797-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0004.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Listen carefully to the entire clip, then report every instance of a plant name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.4, 9.75, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":9.4,\"end\":9.75}]"} {"utterance_id": "8555-284447-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0022.flac", "task_type": "lexical_detection", "query": "mate", "prompt": "Process the full audio offline. Find every instance of the word \"mate\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.58, 1.95, "mate"]], "response_mode": "offline", "answer": "[{\"word\":\"mate\",\"start\":1.58,\"end\":1.95}]"} {"utterance_id": "7127-75947-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0011.flac", "task_type": "lexical_detection", "query": "remain", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"remain\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.35, 0.86, "remain"]], "response_mode": "offline", "answer": "[{\"word\":\"remain\",\"start\":0.35,\"end\":0.86}]"} {"utterance_id": "1284-1180-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0029.flac", "task_type": "lexical_detection", "query": "crazy", "prompt": "When the utterance is complete, report all instances of the word \"crazy\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.51, 1.99, "crazy"]], "response_mode": "offline", "answer": "[{\"word\":\"crazy\",\"start\":1.51,\"end\":1.99}]"} {"utterance_id": "237-126133-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0000.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Do not answer until the audio ends. Then list each instance of a body part as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[5.78, 6.15, "heart"]], "response_mode": "offline", "answer": "[{\"word\":\"heart\",\"start\":5.78,\"end\":6.15}]"} {"utterance_id": "7176-92135-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0024.flac", "task_type": "lexical_detection", "query": "question", "prompt": "Process the full audio offline. Find every instance of the word \"question\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.07, 2.58, "question"]], "response_mode": "offline", "answer": "[{\"word\":\"question\",\"start\":2.07,\"end\":2.58}]"} {"utterance_id": "8455-210777-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0027.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.86, 1.27, "captain"]], "response_mode": "offline", "answer": "[{\"word\":\"captain\",\"start\":0.86,\"end\":1.27}]"} {"utterance_id": "3729-6852-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0043.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.1, 3.34, "douze"]], "response_mode": "offline", "answer": "[{\"word\":\"douze\",\"start\":3.1,\"end\":3.34}]"} {"utterance_id": "6829-68769-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0008.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen carefully to the entire clip, then report every instance of a city name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.2, 0.62, "fairview"]], "response_mode": "offline", "answer": "[{\"word\":\"fairview\",\"start\":0.2,\"end\":0.62}]"} {"utterance_id": "4507-16021-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0030.flac", "task_type": "lexical_counting", "query": "worthy", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"worthy\" (word, start, end). If none, reply [].", "labels": [[5.96, 6.34, "one"], [12.26, 12.75, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":5.96,\"end\":6.34},{\"word\":\"two\",\"start\":12.26,\"end\":12.75}]"} {"utterance_id": "4992-23283-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/23283/4992-23283-0001.flac", "task_type": "lexical_detection", "query": "milner's", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"milner's\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.87, 1.33, "milner's"]], "response_mode": "offline", "answer": "[{\"word\":\"milner's\",\"start\":0.87,\"end\":1.33}]"} {"utterance_id": "7176-88083-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0014.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.55, 1.97, "branch"]], "response_mode": "offline", "answer": "[{\"word\":\"branch\",\"start\":1.55,\"end\":1.97}]"} {"utterance_id": "260-123286-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0014.flac", "task_type": "lexical_detection", "query": "infinite", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"infinite\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.82, 2.31, "infinite"]], "response_mode": "offline", "answer": "[{\"word\":\"infinite\",\"start\":1.82,\"end\":2.31}]"} {"utterance_id": "7729-102255-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0018.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[8.7, 9.27, "officials"]], "response_mode": "offline", "answer": "[{\"word\":\"officials\",\"start\":8.7,\"end\":9.27}]"} {"utterance_id": "260-123286-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0026.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "After listening to the whole recording, output timed count words for an animal name (one, two, three, ...) as JSON. If none, reply [].", "labels": [[0.4, 0.94, "one"], [4.95, 5.67, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.4,\"end\":0.94},{\"word\":\"two\",\"start\":4.95,\"end\":5.67}]"} {"utterance_id": "6829-68771-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0030.flac", "task_type": "lexical_detection", "query": "independent", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"independent\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.21, 1.85, "independent"]], "response_mode": "offline", "answer": "[{\"word\":\"independent\",\"start\":1.21,\"end\":1.85}]"} {"utterance_id": "7176-88083-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0013.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.47, 1.9, "cousin"]], "response_mode": "offline", "answer": "[{\"word\":\"cousin\",\"start\":1.47,\"end\":1.9}]"} {"utterance_id": "4446-2271-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0013.flac", "task_type": "lexical_detection", "query": "conscious", "prompt": "After the audio ends, extract all occurrences of the word \"conscious\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.04, 2.54, "conscious"]], "response_mode": "offline", "answer": "[{\"word\":\"conscious\",\"start\":2.04,\"end\":2.54}]"} {"utterance_id": "7176-92135-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0043.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Once the audio has finished, identify every instance of a number. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.25, 0.4, "two"], [8.49, 8.75, "half"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":0.25,\"end\":0.4},{\"word\":\"half\",\"start\":8.49,\"end\":8.75}]"} {"utterance_id": "61-70968-0035", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0035.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.68, 5.18, "robin's"]], "response_mode": "offline", "answer": "[{\"word\":\"robin's\",\"start\":4.68,\"end\":5.18}]"} {"utterance_id": "6829-68771-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0014.flac", "task_type": "lexical_detection", "query": "louise", "prompt": "After the audio ends, extract all occurrences of the word \"louise\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.01, 3.41, "louise"]], "response_mode": "offline", "answer": "[{\"word\":\"louise\",\"start\":3.01,\"end\":3.41}]"} {"utterance_id": "121-127105-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0007.flac", "task_type": "lexical_detection", "query": "yours", "prompt": "After the audio ends, extract all occurrences of the word \"yours\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.88, 5.45, "yours"]], "response_mode": "offline", "answer": "[{\"word\":\"yours\",\"start\":4.88,\"end\":5.45}]"} {"utterance_id": "3575-170457-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0009.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.75, 6.28, "southey"]], "response_mode": "offline", "answer": "[{\"word\":\"southey\",\"start\":5.75,\"end\":6.28}]"} {"utterance_id": "2094-142345-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0010.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.38, 1.31, "hetty sorrel"]], "response_mode": "offline", "answer": "[{\"word\":\"hetty sorrel\",\"start\":0.38,\"end\":1.31}]"} {"utterance_id": "4992-41797-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0020.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.37, 0.85, "nancy's"], [3.8, 4.24, "olive's"]], "response_mode": "offline", "answer": "[{\"word\":\"nancy's\",\"start\":0.37,\"end\":0.85},{\"word\":\"olive's\",\"start\":3.8,\"end\":4.24}]"} {"utterance_id": "5142-33396-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0011.flac", "task_type": "lexical_detection", "query": "rollers", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"rollers\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.07, 1.52, "rollers"]], "response_mode": "offline", "answer": "[{\"word\":\"rollers\",\"start\":1.07,\"end\":1.52}]"} {"utterance_id": "1188-133604-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0009.flac", "task_type": "lexical_counting", "query": "paper", "prompt": "Listen to the entire clip, then report the count sequence for the word \"paper\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.73, 10.23, "one"], [14.28, 14.65, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":9.73,\"end\":10.23},{\"word\":\"two\",\"start\":14.28,\"end\":14.65}]"} {"utterance_id": "260-123288-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0006.flac", "task_type": "lexical_detection", "query": "charged", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"charged\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.15, 2.79, "charged"]], "response_mode": "offline", "answer": "[{\"word\":\"charged\",\"start\":2.15,\"end\":2.79}]"} {"utterance_id": "61-70968-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0041.flac", "task_type": "lexical_detection", "query": "liked", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"liked\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.53, 3.78, "liked"]], "response_mode": "offline", "answer": "[{\"word\":\"liked\",\"start\":3.53,\"end\":3.78}]"} {"utterance_id": "4970-29095-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0038.flac", "task_type": "lexical_detection", "query": "push", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"push\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.96, 3.23, "push"]], "response_mode": "offline", "answer": "[{\"word\":\"push\",\"start\":2.96,\"end\":3.23}]"} {"utterance_id": "7176-92135-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0041.flac", "task_type": "semantic_detection", "query": "food", "prompt": "When the utterance is complete, report all instances of a food name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.91, 3.12, "tea"]], "response_mode": "offline", "answer": "[{\"word\":\"tea\",\"start\":2.91,\"end\":3.12}]"} {"utterance_id": "4446-2275-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0043.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.49, 0.84, "bartley"]], "response_mode": "offline", "answer": "[{\"word\":\"bartley\",\"start\":0.49,\"end\":0.84}]"} {"utterance_id": "2961-961-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0009.flac", "task_type": "lexical_counting", "query": "story", "prompt": "Once the audio has finished, list each occurrence of the word \"story\" as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[2.6, 3.26, "one"], [4.78, 5.28, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.6,\"end\":3.26},{\"word\":\"two\",\"start\":4.78,\"end\":5.28}]"} {"utterance_id": "6829-68771-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0023.flac", "task_type": "lexical_detection", "query": "educated", "prompt": "Process the full audio offline. Find every instance of the word \"educated\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.35, 1.9, "educated"]], "response_mode": "offline", "answer": "[{\"word\":\"educated\",\"start\":1.35,\"end\":1.9}]"} {"utterance_id": "1995-1837-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0005.flac", "task_type": "lexical_detection", "query": "creature", "prompt": "After the audio ends, extract all occurrences of the word \"creature\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.76, 2.24, "creature"]], "response_mode": "offline", "answer": "[{\"word\":\"creature\",\"start\":1.76,\"end\":2.24}]"} {"utterance_id": "1580-141084-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0016.flac", "task_type": "lexical_detection", "query": "humorous", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"humorous\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.32, 4.79, "humorous"]], "response_mode": "offline", "answer": "[{\"word\":\"humorous\",\"start\":4.32,\"end\":4.79}]"} {"utterance_id": "7127-75947-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0037.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the entire clip, then report the count sequence for a person's name as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.47, 3.11, "one"], [7.88, 8.34, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.47,\"end\":3.11},{\"word\":\"two\",\"start\":7.88,\"end\":8.34}]"} {"utterance_id": "2300-131720-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0019.flac", "task_type": "lexical_counting", "query": "system", "prompt": "After listening to the whole recording, output timed count words for the word \"system\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[2.48, 3.07, "one"], [10.39, 10.85, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.48,\"end\":3.07},{\"word\":\"two\",\"start\":10.39,\"end\":10.85}]"} {"utterance_id": "4992-41797-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0012.flac", "task_type": "lexical_detection", "query": "wild", "prompt": "Once the audio has finished, identify every instance of the word \"wild\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.68, 1.28, "wild"]], "response_mode": "offline", "answer": "[{\"word\":\"wild\",\"start\":0.68,\"end\":1.28}]"} {"utterance_id": "1221-135767-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0023.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[8.73, 9.13, "settler"]], "response_mode": "offline", "answer": "[{\"word\":\"settler\",\"start\":8.73,\"end\":9.13}]"} {"utterance_id": "1221-135767-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0023.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Process the full audio offline. Find every instance of a plant name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.0, 1.7, "rose bushes"], [3.1, 3.85, "apple trees"]], "response_mode": "offline", "answer": "[{\"word\":\"rose bushes\",\"start\":1.0,\"end\":1.7},{\"word\":\"apple trees\",\"start\":3.1,\"end\":3.85}]"} {"utterance_id": "4970-29095-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0005.flac", "task_type": "lexical_counting", "query": "thy", "prompt": "After the audio ends, tally every instance of the word \"thy\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.55, 0.86, "one"], [3.5, 3.66, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.55,\"end\":0.86},{\"word\":\"two\",\"start\":3.5,\"end\":3.66}]"} {"utterance_id": "1320-122617-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0031.flac", "task_type": "lexical_detection", "query": "dare", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"dare\" with fields word, start, end (seconds). If none, reply [].", "labels": [[4.49, 4.86, "dare"]], "response_mode": "offline", "answer": "[{\"word\":\"dare\",\"start\":4.49,\"end\":4.86}]"} {"utterance_id": "6930-76324-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0015.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a time of day with fields word, start, end (seconds). If none, reply [].", "labels": [[5.19, 5.9, "afternoon"]], "response_mode": "offline", "answer": "[{\"word\":\"afternoon\",\"start\":5.19,\"end\":5.9}]"} {"utterance_id": "2300-131720-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0031.flac", "task_type": "lexical_detection", "query": "details", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"details\" with fields word, start, end (seconds). If none, reply [].", "labels": [[5.42, 6.21, "details"]], "response_mode": "offline", "answer": "[{\"word\":\"details\",\"start\":5.42,\"end\":6.21}]"} {"utterance_id": "3575-170457-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0000.flac", "task_type": "lexical_detection", "query": "feared", "prompt": "Once the audio has finished, identify every instance of the word \"feared\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.83, 5.17, "feared"]], "response_mode": "offline", "answer": "[{\"word\":\"feared\",\"start\":4.83,\"end\":5.17}]"} {"utterance_id": "8555-292519-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0013.flac", "task_type": "lexical_detection", "query": "rustling", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"rustling\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.27, 1.84, "rustling"]], "response_mode": "offline", "answer": "[{\"word\":\"rustling\",\"start\":1.27,\"end\":1.84}]"} {"utterance_id": "8555-284447-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0003.flac", "task_type": "lexical_detection", "query": "cap'n", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"cap'n\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.48, 0.78, "cap'n"]], "response_mode": "offline", "answer": "[{\"word\":\"cap'n\",\"start\":0.48,\"end\":0.78}]"} {"utterance_id": "2300-131720-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0022.flac", "task_type": "lexical_detection", "query": "report", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"report\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.97, 3.45, "report"]], "response_mode": "offline", "answer": "[{\"word\":\"report\",\"start\":2.97,\"end\":3.45}]"} {"utterance_id": "8455-210777-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0016.flac", "task_type": "lexical_detection", "query": "ordinary", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"ordinary\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[14.84, 15.39, "ordinary"]], "response_mode": "offline", "answer": "[{\"word\":\"ordinary\",\"start\":14.84,\"end\":15.39}]"} {"utterance_id": "6829-68769-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0031.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[4.92, 5.14, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.92,\"end\":5.14}]"} {"utterance_id": "8224-274384-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8224/274384/8224-274384-0004.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a profession (word, start, end). If none, reply [].", "labels": [[6.4, 6.87, "one"], [19.57, 20.14, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.4,\"end\":6.87},{\"word\":\"two\",\"start\":19.57,\"end\":20.14}]"} {"utterance_id": "121-127105-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0022.flac", "task_type": "lexical_counting", "query": "know", "prompt": "Listen to the entire clip, then report the count sequence for the word \"know\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.68, 1.86, "one"], [3.66, 3.89, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.68,\"end\":1.86},{\"word\":\"two\",\"start\":3.66,\"end\":3.89}]"} {"utterance_id": "1089-134686-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0028.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen to the full audio. After it finishes, find every instance of a time of day. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.47, 3.21, "afternoon"]], "response_mode": "offline", "answer": "[{\"word\":\"afternoon\",\"start\":2.47,\"end\":3.21}]"} {"utterance_id": "3570-5695-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5695/3570-5695-0003.flac", "task_type": "lexical_detection", "query": "culture", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"culture\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.6, 2.1, "culture"]], "response_mode": "offline", "answer": "[{\"word\":\"culture\",\"start\":1.6,\"end\":2.1}]"} {"utterance_id": "5142-36377-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0019.flac", "task_type": "lexical_detection", "query": "addressed", "prompt": "When the utterance is complete, report all instances of the word \"addressed\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.46, 0.91, "addressed"]], "response_mode": "offline", "answer": "[{\"word\":\"addressed\",\"start\":0.46,\"end\":0.91}]"} {"utterance_id": "1995-1826-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1826/1995-1826-0021.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.51, 4.18, "miss taylor"]], "response_mode": "offline", "answer": "[{\"word\":\"miss taylor\",\"start\":3.51,\"end\":4.18}]"} {"utterance_id": "8230-279154-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0002.flac", "task_type": "lexical_detection", "query": "agree", "prompt": "Process the full audio offline. Find every instance of the word \"agree\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.14, 4.46, "agree"]], "response_mode": "offline", "answer": "[{\"word\":\"agree\",\"start\":4.14,\"end\":4.46}]"} {"utterance_id": "3729-6852-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0045.flac", "task_type": "lexical_detection", "query": "surrounded", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"surrounded\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[15.17, 15.69, "surrounded"]], "response_mode": "offline", "answer": "[{\"word\":\"surrounded\",\"start\":15.17,\"end\":15.69}]"} {"utterance_id": "237-126133-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0022.flac", "task_type": "lexical_detection", "query": "fears", "prompt": "Process the full audio offline. Find every instance of the word \"fears\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.33, 1.74, "fears"]], "response_mode": "offline", "answer": "[{\"word\":\"fears\",\"start\":1.33,\"end\":1.74}]"} {"utterance_id": "1089-134691-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0012.flac", "task_type": "lexical_detection", "query": "beaten", "prompt": "When the utterance is complete, report all instances of the word \"beaten\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.66, 9.16, "beaten"]], "response_mode": "offline", "answer": "[{\"word\":\"beaten\",\"start\":8.66,\"end\":9.16}]"} {"utterance_id": "4507-16021-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0034.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the full audio. After it finishes, find every instance of a number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.28, 1.65, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":1.28,\"end\":1.65}]"} {"utterance_id": "61-70968-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0017.flac", "task_type": "lexical_detection", "query": "injured", "prompt": "Process the full audio offline. Find every instance of the word \"injured\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.66, 2.04, "injured"]], "response_mode": "offline", "answer": "[{\"word\":\"injured\",\"start\":1.66,\"end\":2.04}]"} {"utterance_id": "2961-961-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0022.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.42, 6.22, "timaeus"]], "response_mode": "offline", "answer": "[{\"word\":\"timaeus\",\"start\":5.42,\"end\":6.22}]"} {"utterance_id": "1580-141084-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0039.flac", "task_type": "lexical_detection", "query": "confidence", "prompt": "Once the audio has finished, identify every instance of the word \"confidence\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.14, 2.72, "confidence"]], "response_mode": "offline", "answer": "[{\"word\":\"confidence\",\"start\":2.14,\"end\":2.72}]"} {"utterance_id": "2830-3980-0045", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0045.flac", "task_type": "lexical_detection", "query": "nature", "prompt": "Once the audio has finished, identify every instance of the word \"nature\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.21, 2.59, "nature"]], "response_mode": "offline", "answer": "[{\"word\":\"nature\",\"start\":2.21,\"end\":2.59}]"} {"utterance_id": "1221-135767-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0003.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Once the audio has finished, identify every instance of an organization name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[10.73, 11.88, "legislative body"], [17.3, 18.12, "legislature"]], "response_mode": "offline", "answer": "[{\"word\":\"legislative body\",\"start\":10.73,\"end\":11.88},{\"word\":\"legislature\",\"start\":17.3,\"end\":18.12}]"} {"utterance_id": "4992-41806-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0003.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[0.54, 1.13, "kathleen"]], "response_mode": "offline", "answer": "[{\"word\":\"kathleen\",\"start\":0.54,\"end\":1.13}]"} {"utterance_id": "260-123440-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0001.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.92, 1.56, "alice"]], "response_mode": "offline", "answer": "[{\"word\":\"alice\",\"start\":0.92,\"end\":1.56}]"} {"utterance_id": "4446-2273-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0006.flac", "task_type": "lexical_detection", "query": "made", "prompt": "When the utterance is complete, report all instances of the word \"made\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.1, 1.32, "made"]], "response_mode": "offline", "answer": "[{\"word\":\"made\",\"start\":1.1,\"end\":1.32}]"} {"utterance_id": "6829-68771-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0001.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Listen carefully to the entire clip, then report every instance of a state or province name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.45, 7.93, "district"]], "response_mode": "offline", "answer": "[{\"word\":\"district\",\"start\":7.45,\"end\":7.93}]"} {"utterance_id": "3575-170457-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0023.flac", "task_type": "lexical_detection", "query": "amongst", "prompt": "When the utterance is complete, report all instances of the word \"amongst\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.33, 6.74, "amongst"]], "response_mode": "offline", "answer": "[{\"word\":\"amongst\",\"start\":6.33,\"end\":6.74}]"} {"utterance_id": "1995-1836-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0002.flac", "task_type": "lexical_detection", "query": "why", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"why\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.15, 0.4, "why"]], "response_mode": "offline", "answer": "[{\"word\":\"why\",\"start\":0.15,\"end\":0.4}]"} {"utterance_id": "908-31957-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0011.flac", "task_type": "lexical_detection", "query": "false", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"false\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.15, 2.05, "false"]], "response_mode": "offline", "answer": "[{\"word\":\"false\",\"start\":1.15,\"end\":2.05}]"} {"utterance_id": "7176-92135-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0006.flac", "task_type": "lexical_detection", "query": "i'd", "prompt": "Once the audio has finished, identify every instance of the word \"i'd\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.73, 5.81, "i'd"]], "response_mode": "offline", "answer": "[{\"word\":\"i'd\",\"start\":5.73,\"end\":5.81}]"} {"utterance_id": "121-127105-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/127105/121-127105-0023.flac", "task_type": "lexical_detection", "query": "narrative", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"narrative\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.59, 4.38, "narrative"]], "response_mode": "offline", "answer": "[{\"word\":\"narrative\",\"start\":3.59,\"end\":4.38}]"} {"utterance_id": "2300-131720-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0033.flac", "task_type": "lexical_detection", "query": "eighty", "prompt": "Process the full audio offline. Find every instance of the word \"eighty\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.69, 1.98, "eighty"]], "response_mode": "offline", "answer": "[{\"word\":\"eighty\",\"start\":1.69,\"end\":1.98}]"} {"utterance_id": "4507-16021-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0029.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a number with fields word, start, end (seconds). If none, reply [].", "labels": [[2.1, 2.46, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.1,\"end\":2.46}]"} {"utterance_id": "4446-2275-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0021.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.53, 0.93, "hilda's"]], "response_mode": "offline", "answer": "[{\"word\":\"hilda's\",\"start\":0.53,\"end\":0.93}]"} {"utterance_id": "5142-33396-0064", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0064.flac", "task_type": "lexical_detection", "query": "looked", "prompt": "When the utterance is complete, report all instances of the word \"looked\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.69, 1.88, "looked"]], "response_mode": "offline", "answer": "[{\"word\":\"looked\",\"start\":1.69,\"end\":1.88}]"} {"utterance_id": "6829-68769-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0022.flac", "task_type": "lexical_detection", "query": "interested", "prompt": "Once the audio has finished, identify every instance of the word \"interested\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.18, 3.69, "interested"]], "response_mode": "offline", "answer": "[{\"word\":\"interested\",\"start\":3.18,\"end\":3.69}]"} {"utterance_id": "2830-3980-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0006.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.11, 2.72, "apostles"]], "response_mode": "offline", "answer": "[{\"word\":\"apostles\",\"start\":2.11,\"end\":2.72}]"} {"utterance_id": "7021-85628-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0024.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.23, 0.67, "anders"]], "response_mode": "offline", "answer": "[{\"word\":\"anders\",\"start\":0.23,\"end\":0.67}]"} {"utterance_id": "1320-122617-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0040.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Do not answer until the audio ends. Then list each instance of a city name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[7.14, 7.64, "village"]], "response_mode": "offline", "answer": "[{\"word\":\"village\",\"start\":7.14,\"end\":7.64}]"} {"utterance_id": "672-122797-0055", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0055.flac", "task_type": "lexical_detection", "query": "listened", "prompt": "Process the full audio offline. Find every instance of the word \"listened\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.82, 7.42, "listened"]], "response_mode": "offline", "answer": "[{\"word\":\"listened\",\"start\":6.82,\"end\":7.42}]"} {"utterance_id": "260-123440-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123440/260-123440-0002.flac", "task_type": "semantic_counting", "query": "cardinal", "prompt": "Once the audio has finished, list each occurrence of a number as {\"word\": , \"start\", \"end\"} in order. If none, reply [].", "labels": [[3.37, 3.62, "one"], [4.99, 5.17, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.37,\"end\":3.62},{\"word\":\"two\",\"start\":4.99,\"end\":5.17}]"} {"utterance_id": "121-121726-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0003.flac", "task_type": "lexical_detection", "query": "fever", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"fever\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.05, 1.67, "fever"]], "response_mode": "offline", "answer": "[{\"word\":\"fever\",\"start\":1.05,\"end\":1.67}]"} {"utterance_id": "3570-5696-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0005.flac", "task_type": "semantic_detection", "query": "money", "prompt": "You will hear a complete utterance. When it ends, list every instance of a money or currency term you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.38, 3.13, "expenditure"], [11.13, 11.79, "expenditure"]], "response_mode": "offline", "answer": "[{\"word\":\"expenditure\",\"start\":2.38,\"end\":3.13},{\"word\":\"expenditure\",\"start\":11.13,\"end\":11.79}]"} {"utterance_id": "61-70970-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0005.flac", "task_type": "lexical_detection", "query": "lad", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"lad\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.83, "lad"]], "response_mode": "offline", "answer": "[{\"word\":\"lad\",\"start\":0.53,\"end\":0.83}]"} {"utterance_id": "5142-33396-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0046.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.91, 1.26, "odin"]], "response_mode": "offline", "answer": "[{\"word\":\"odin\",\"start\":0.91,\"end\":1.26}]"} {"utterance_id": "3575-170457-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0002.flac", "task_type": "lexical_detection", "query": "divided", "prompt": "When the utterance is complete, report all instances of the word \"divided\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.32, 1.99, "divided"]], "response_mode": "offline", "answer": "[{\"word\":\"divided\",\"start\":1.32,\"end\":1.99}]"} {"utterance_id": "1995-1836-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1836/1995-1836-0001.flac", "task_type": "lexical_detection", "query": "senate", "prompt": "Process the full audio offline. Find every instance of the word \"senate\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.22, 5.72, "senate"]], "response_mode": "offline", "answer": "[{\"word\":\"senate\",\"start\":5.22,\"end\":5.72}]"} {"utterance_id": "1089-134691-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0002.flac", "task_type": "lexical_detection", "query": "abruptly", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"abruptly\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.19, 1.71, "abruptly"]], "response_mode": "offline", "answer": "[{\"word\":\"abruptly\",\"start\":1.19,\"end\":1.71}]"} {"utterance_id": "5142-33396-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0029.flac", "task_type": "lexical_detection", "query": "bring", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"bring\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.2, 0.45, "bring"]], "response_mode": "offline", "answer": "[{\"word\":\"bring\",\"start\":0.2,\"end\":0.45}]"} {"utterance_id": "3575-170457-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0007.flac", "task_type": "lexical_detection", "query": "yorkshire", "prompt": "Process the full audio offline. Find every instance of the word \"yorkshire\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.72, 7.33, "yorkshire"]], "response_mode": "offline", "answer": "[{\"word\":\"yorkshire\",\"start\":6.72,\"end\":7.33}]"} {"utterance_id": "260-123288-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123288/260-123288-0002.flac", "task_type": "lexical_detection", "query": "atmosphere", "prompt": "When the utterance is complete, report all instances of the word \"atmosphere\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.58, 1.14, "atmosphere"]], "response_mode": "offline", "answer": "[{\"word\":\"atmosphere\",\"start\":0.58,\"end\":1.14}]"} {"utterance_id": "237-134500-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0031.flac", "task_type": "lexical_detection", "query": "corners", "prompt": "Once the audio has finished, identify every instance of the word \"corners\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.29, 4.7, "corners"]], "response_mode": "offline", "answer": "[{\"word\":\"corners\",\"start\":4.29,\"end\":4.7}]"} {"utterance_id": "6829-68769-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0012.flac", "task_type": "lexical_detection", "query": "observed", "prompt": "After the audio ends, extract all occurrences of the word \"observed\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.9, 2.44, "observed"]], "response_mode": "offline", "answer": "[{\"word\":\"observed\",\"start\":1.9,\"end\":2.44}]"} {"utterance_id": "7176-88083-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0013.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After the audio ends, extract all occurrences of a plant name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.05, 11.57, "weeds"]], "response_mode": "offline", "answer": "[{\"word\":\"weeds\",\"start\":11.05,\"end\":11.57}]"} {"utterance_id": "908-31957-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0018.flac", "task_type": "lexical_detection", "query": "beloved", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"beloved\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[3.12, 3.7, "beloved"]], "response_mode": "offline", "answer": "[{\"word\":\"beloved\",\"start\":3.12,\"end\":3.7}]"} {"utterance_id": "5142-36377-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0008.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After listening to the whole recording, output every instance of a number with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.38, 2.6, "one"], [14.47, 14.68, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.38,\"end\":2.6},{\"word\":\"two\",\"start\":14.47,\"end\":14.68}]"} {"utterance_id": "3729-6852-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0006.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.43, 0.94, "silvia"]], "response_mode": "offline", "answer": "[{\"word\":\"silvia\",\"start\":0.43,\"end\":0.94}]"} {"utterance_id": "1221-135767-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0012.flac", "task_type": "lexical_detection", "query": "wooden", "prompt": "Once the audio has finished, identify every instance of the word \"wooden\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[10.98, 11.28, "wooden"]], "response_mode": "offline", "answer": "[{\"word\":\"wooden\",\"start\":10.98,\"end\":11.28}]"} {"utterance_id": "2094-142345-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0054.flac", "task_type": "lexical_detection", "query": "wouldn't", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"wouldn't\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.36, 4.63, "wouldn't"]], "response_mode": "offline", "answer": "[{\"word\":\"wouldn't\",\"start\":4.36,\"end\":4.63}]"} {"utterance_id": "1320-122612-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.28, 0.98, "chingachgook"]], "response_mode": "offline", "answer": "[{\"word\":\"chingachgook\",\"start\":0.28,\"end\":0.98}]"} {"utterance_id": "4970-29093-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29093/4970-29093-0002.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.91, 7.28, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":6.91,\"end\":7.28}]"} {"utterance_id": "1188-133604-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0041.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "When the utterance is complete, report all instances of a plant name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.05, 3.56, "corn"]], "response_mode": "offline", "answer": "[{\"word\":\"corn\",\"start\":3.05,\"end\":3.56}]"} {"utterance_id": "672-122797-0064", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0064.flac", "task_type": "lexical_counting", "query": "little", "prompt": "After the audio ends, tally every instance of the word \"little\". Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[1.55, 1.79, "one"], [11.07, 11.32, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.55,\"end\":1.79},{\"word\":\"two\",\"start\":11.07,\"end\":11.32}]"} {"utterance_id": "5105-28241-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0003.flac", "task_type": "lexical_detection", "query": "canvas", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"canvas\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.32, 1.79, "canvas"]], "response_mode": "offline", "answer": "[{\"word\":\"canvas\",\"start\":1.32,\"end\":1.79}]"} {"utterance_id": "6930-76324-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/76324/6930-76324-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.44, 2.46, "cynthia sprague"]], "response_mode": "offline", "answer": "[{\"word\":\"cynthia sprague\",\"start\":1.44,\"end\":2.46}]"} {"utterance_id": "4446-2275-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0018.flac", "task_type": "lexical_detection", "query": "misery", "prompt": "After listening to the whole recording, output every instance of the word \"misery\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.0, 1.44, "misery"]], "response_mode": "offline", "answer": "[{\"word\":\"misery\",\"start\":1.0,\"end\":1.44}]"} {"utterance_id": "5683-32866-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0018.flac", "task_type": "lexical_detection", "query": "awful", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"awful\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.6, 4.13, "awful"]], "response_mode": "offline", "answer": "[{\"word\":\"awful\",\"start\":3.6,\"end\":4.13}]"} {"utterance_id": "4970-29095-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0020.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a body part with fields word, start, end (seconds). If none, reply [].", "labels": [[10.87, 11.15, "eye"]], "response_mode": "offline", "answer": "[{\"word\":\"eye\",\"start\":10.87,\"end\":11.15}]"} {"utterance_id": "2830-3980-0040", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0040.flac", "task_type": "lexical_detection", "query": "greeting", "prompt": "Process the full audio offline. Find every instance of the word \"greeting\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.66, 0.98, "greeting"]], "response_mode": "offline", "answer": "[{\"word\":\"greeting\",\"start\":0.66,\"end\":0.98}]"} {"utterance_id": "6930-75918-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0004.flac", "task_type": "lexical_detection", "query": "brilliant", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"brilliant\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.58, 9.32, "brilliant"]], "response_mode": "offline", "answer": "[{\"word\":\"brilliant\",\"start\":8.58,\"end\":9.32}]"} {"utterance_id": "1089-134691-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134691/1089-134691-0011.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "After the audio ends, extract all occurrences of a number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[16.04, 16.8, "tenfold"]], "response_mode": "offline", "answer": "[{\"word\":\"tenfold\",\"start\":16.04,\"end\":16.8}]"} {"utterance_id": "2300-131720-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0013.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Do not answer until the audio ends. Then list each instance of a person's name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[9.0, 9.51, "edison"]], "response_mode": "offline", "answer": "[{\"word\":\"edison\",\"start\":9.0,\"end\":9.51}]"} {"utterance_id": "1320-122617-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0025.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Listen to the full audio. After it finishes, find every instance of a body part. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.32, 3.63, "skin"]], "response_mode": "offline", "answer": "[{\"word\":\"skin\",\"start\":3.32,\"end\":3.63}]"} {"utterance_id": "2961-961-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0010.flac", "task_type": "semantic_detection", "query": "country", "prompt": "Process the full audio offline. Find every instance of a country name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.75, 1.47, "egypt"]], "response_mode": "offline", "answer": "[{\"word\":\"egypt\",\"start\":0.75,\"end\":1.47}]"} {"utterance_id": "1089-134686-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0026.flac", "task_type": "lexical_detection", "query": "rector", "prompt": "After the audio ends, extract all occurrences of the word \"rector\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.65, 1.0, "rector"]], "response_mode": "offline", "answer": "[{\"word\":\"rector\",\"start\":0.65,\"end\":1.0}]"} {"utterance_id": "61-70968-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0041.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.04, 4.26, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":4.04,\"end\":4.26}]"} {"utterance_id": "908-31957-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0006.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Process the full audio offline. Find every instance of an animal name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.92, 5.42, "dove"]], "response_mode": "offline", "answer": "[{\"word\":\"dove\",\"start\":4.92,\"end\":5.42}]"} {"utterance_id": "3729-6852-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0017.flac", "task_type": "lexical_counting", "query": "under", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"under\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[6.27, 6.45, "one"], [16.46, 16.66, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":6.27,\"end\":6.45},{\"word\":\"two\",\"start\":16.46,\"end\":16.66}]"} {"utterance_id": "4077-13751-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0012.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.54, 3.11, "lilburn w boggs"]], "response_mode": "offline", "answer": "[{\"word\":\"lilburn w boggs\",\"start\":1.54,\"end\":3.11}]"} {"utterance_id": "1580-141083-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0041.flac", "task_type": "lexical_detection", "query": "suspicions", "prompt": "When the utterance is complete, report all instances of the word \"suspicions\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.95, 1.66, "suspicions"]], "response_mode": "offline", "answer": "[{\"word\":\"suspicions\",\"start\":0.95,\"end\":1.66}]"} {"utterance_id": "260-123286-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0011.flac", "task_type": "semantic_detection", "query": "weather", "prompt": "After the audio ends, extract all occurrences of a weather term. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.23, 1.56, "weather"], [2.92, 3.25, "wind"]], "response_mode": "offline", "answer": "[{\"word\":\"weather\",\"start\":1.23,\"end\":1.56},{\"word\":\"wind\",\"start\":2.92,\"end\":3.25}]"} {"utterance_id": "1188-133604-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0012.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "Listen to the full audio. Count every occurrence of a profession. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.51, 2.36, "one"], [5.2, 5.57, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.51,\"end\":2.36},{\"word\":\"two\",\"start\":5.2,\"end\":5.57}]"} {"utterance_id": "7127-75947-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0029.flac", "task_type": "lexical_counting", "query": "indeed", "prompt": "Process the full audio offline. For every occurrence of the word \"indeed\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.39, 2.01, "one"], [3.58, 4.13, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.39,\"end\":2.01},{\"word\":\"two\",\"start\":3.58,\"end\":4.13}]"} {"utterance_id": "1221-135766-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135766/1221-135766-0011.flac", "task_type": "semantic_counting", "query": "body_part", "prompt": "Process the full audio offline. For every occurrence of a body part, emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[10.55, 11.02, "one"], [18.1, 18.4, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":10.55,\"end\":11.02},{\"word\":\"two\",\"start\":18.1,\"end\":18.4}]"} {"utterance_id": "2300-131720-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0017.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[15.8, 15.97, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":15.8,\"end\":15.97}]"} {"utterance_id": "1188-133604-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0006.flac", "task_type": "lexical_detection", "query": "beak", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"beak\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.39, 1.66, "beak"]], "response_mode": "offline", "answer": "[{\"word\":\"beak\",\"start\":1.39,\"end\":1.66}]"} {"utterance_id": "4507-16021-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0017.flac", "task_type": "lexical_counting", "query": "fact", "prompt": "When the utterance is complete, report how the word \"fact\" accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.12, 4.56, "one"], [8.77, 9.24, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.12,\"end\":4.56},{\"word\":\"two\",\"start\":8.77,\"end\":9.24}]"} {"utterance_id": "260-123286-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0018.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Process the full audio offline. Find every instance of an organization name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.72, 2.33, "museum"]], "response_mode": "offline", "answer": "[{\"word\":\"museum\",\"start\":1.72,\"end\":2.33}]"} {"utterance_id": "6930-75918-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0001.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "When the utterance is complete, report all instances of an ordinal number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.26, 13.64, "next"]], "response_mode": "offline", "answer": "[{\"word\":\"next\",\"start\":13.26,\"end\":13.64}]"} {"utterance_id": "8463-294828-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0036.flac", "task_type": "semantic_counting", "query": "city", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a city name: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[1.27, 1.78, "one"], [2.83, 3.3, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.27,\"end\":1.78},{\"word\":\"two\",\"start\":2.83,\"end\":3.3}]"} {"utterance_id": "3729-6852-0031", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0031.flac", "task_type": "lexical_detection", "query": "render", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"render\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.42, 2.76, "render"]], "response_mode": "offline", "answer": "[{\"word\":\"render\",\"start\":2.42,\"end\":2.76}]"} {"utterance_id": "672-122797-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0020.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Once the audio has finished, identify every instance of a plant name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.83, 1.39, "tree"]], "response_mode": "offline", "answer": "[{\"word\":\"tree\",\"start\":0.83,\"end\":1.39}]"} {"utterance_id": "8463-294828-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0015.flac", "task_type": "lexical_counting", "query": "conseil", "prompt": "Process the full audio offline. For every occurrence of the word \"conseil\", emit a count-word event with timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 0.93, "one"], [3.38, 3.95, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.53,\"end\":0.93},{\"word\":\"two\",\"start\":3.38,\"end\":3.95}]"} {"utterance_id": "672-122797-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0028.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a time of day with fields word, start, end (seconds). If none, reply [].", "labels": [[0.93, 1.54, "evening"]], "response_mode": "offline", "answer": "[{\"word\":\"evening\",\"start\":0.93,\"end\":1.54}]"} {"utterance_id": "1580-141084-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141084/1580-141084-0027.flac", "task_type": "lexical_detection", "query": "certainly", "prompt": "When the utterance is complete, report all instances of the word \"certainly\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.56, 2.31, "certainly"]], "response_mode": "offline", "answer": "[{\"word\":\"certainly\",\"start\":1.56,\"end\":2.31}]"} {"utterance_id": "4077-13754-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0009.flac", "task_type": "lexical_detection", "query": "plural", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"plural\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.41, 1.71, "plural"]], "response_mode": "offline", "answer": "[{\"word\":\"plural\",\"start\":1.41,\"end\":1.71}]"} {"utterance_id": "2094-142345-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0013.flac", "task_type": "lexical_detection", "query": "unfinished", "prompt": "Once the audio has finished, identify every instance of the word \"unfinished\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[7.1, 7.74, "unfinished"]], "response_mode": "offline", "answer": "[{\"word\":\"unfinished\",\"start\":7.1,\"end\":7.74}]"} {"utterance_id": "237-134500-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0036.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.0, 6.38, "marie"]], "response_mode": "offline", "answer": "[{\"word\":\"marie\",\"start\":6.0,\"end\":6.38}]"} {"utterance_id": "1284-1180-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0026.flac", "task_type": "lexical_detection", "query": "cook", "prompt": "Once the audio has finished, identify every instance of the word \"cook\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[6.28, 6.51, "cook"]], "response_mode": "offline", "answer": "[{\"word\":\"cook\",\"start\":6.28,\"end\":6.51}]"} {"utterance_id": "260-123286-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0020.flac", "task_type": "semantic_detection", "query": "date", "prompt": "Listen carefully to the entire clip, then report every instance of a date as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 2.6, "tuesday august eighteenth"]], "response_mode": "offline", "answer": "[{\"word\":\"tuesday august eighteenth\",\"start\":0.53,\"end\":2.6}]"} {"utterance_id": "5683-32865-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0003.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After the audio ends, extract all occurrences of a family relation. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.66, 1.22, "cousins"], [2.61, 3.33, "cousins"]], "response_mode": "offline", "answer": "[{\"word\":\"cousins\",\"start\":0.66,\"end\":1.22},{\"word\":\"cousins\",\"start\":2.61,\"end\":3.33}]"} {"utterance_id": "3729-6852-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0017.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After listening to the whole recording, output every instance of a plant name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.12, 7.73, "trees"], [16.79, 17.47, "foliage"]], "response_mode": "offline", "answer": "[{\"word\":\"trees\",\"start\":7.12,\"end\":7.73},{\"word\":\"foliage\",\"start\":16.79,\"end\":17.47}]"} {"utterance_id": "7127-75947-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0015.flac", "task_type": "lexical_detection", "query": "kindly", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"kindly\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.27, 2.86, "kindly"]], "response_mode": "offline", "answer": "[{\"word\":\"kindly\",\"start\":2.27,\"end\":2.86}]"} {"utterance_id": "4077-13751-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0018.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Once the audio has finished, identify every instance of an animal name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[13.67, 14.22, "beasts"]], "response_mode": "offline", "answer": "[{\"word\":\"beasts\",\"start\":13.67,\"end\":14.22}]"} {"utterance_id": "3729-6852-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0029.flac", "task_type": "lexical_counting", "query": "sold", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"sold\" with word/start/end. If none, reply [].", "labels": [[0.86, 1.15, "one"], [6.23, 6.75, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.86,\"end\":1.15},{\"word\":\"two\",\"start\":6.23,\"end\":6.75}]"} {"utterance_id": "8463-294825-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0017.flac", "task_type": "lexical_detection", "query": "roughly", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"roughly\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.9, 1.31, "roughly"]], "response_mode": "offline", "answer": "[{\"word\":\"roughly\",\"start\":0.9,\"end\":1.31}]"} {"utterance_id": "7021-85628-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0010.flac", "task_type": "semantic_counting", "query": "color", "prompt": "After the audio ends, tally every instance of a color name. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.9, 1.26, "one"], [4.96, 5.29, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.9,\"end\":1.26},{\"word\":\"two\",\"start\":4.96,\"end\":5.29}]"} {"utterance_id": "3570-5694-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5694/3570-5694-0011.flac", "task_type": "lexical_counting", "query": "motives", "prompt": "Listen to the full audio. Count every occurrence of the word \"motives\". After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.91, 3.36, "one"], [5.73, 6.21, "two"], [10.21, 10.8, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":2.91,\"end\":3.36},{\"word\":\"two\",\"start\":5.73,\"end\":6.21},{\"word\":\"three\",\"start\":10.21,\"end\":10.8}]"} {"utterance_id": "8555-284447-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0007.flac", "task_type": "lexical_detection", "query": "majesty", "prompt": "After the audio ends, extract all occurrences of the word \"majesty\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.02, 1.56, "majesty"]], "response_mode": "offline", "answer": "[{\"word\":\"majesty\",\"start\":1.02,\"end\":1.56}]"} {"utterance_id": "5683-32879-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32879/5683-32879-0000.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Once the audio has finished, identify every instance of an animal name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.36, 3.71, "pony"]], "response_mode": "offline", "answer": "[{\"word\":\"pony\",\"start\":3.36,\"end\":3.71}]"} {"utterance_id": "908-31957-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0017.flac", "task_type": "lexical_detection", "query": "smile", "prompt": "Once the audio has finished, identify every instance of the word \"smile\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.22, 4.86, "smile"]], "response_mode": "offline", "answer": "[{\"word\":\"smile\",\"start\":4.22,\"end\":4.86}]"} {"utterance_id": "3575-170457-0054", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0054.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Do not answer until the audio ends. Then list each instance of a family relation as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.42, 2.96, "sister"]], "response_mode": "offline", "answer": "[{\"word\":\"sister\",\"start\":2.42,\"end\":2.96}]"} {"utterance_id": "2830-3980-0057", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0057.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, produce a running count of a person's name: each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[1.31, 1.67, "one"], [6.2, 6.46, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.31,\"end\":1.67},{\"word\":\"two\",\"start\":6.2,\"end\":6.46}]"} {"utterance_id": "5142-33396-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0016.flac", "task_type": "semantic_detection", "query": "country", "prompt": "After the audio ends, extract all occurrences of a country name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.45, 2.03, "norway"]], "response_mode": "offline", "answer": "[{\"word\":\"norway\",\"start\":1.45,\"end\":2.03}]"} {"utterance_id": "5142-33396-0067", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0067.flac", "task_type": "lexical_detection", "query": "sharp", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"sharp\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.61, 1.0, "sharp"]], "response_mode": "offline", "answer": "[{\"word\":\"sharp\",\"start\":0.61,\"end\":1.0}]"} {"utterance_id": "237-126133-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0008.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.65, 2.07, "face"]], "response_mode": "offline", "answer": "[{\"word\":\"face\",\"start\":1.65,\"end\":2.07}]"} {"utterance_id": "3575-170457-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0007.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a state or province name with fields word, start, end (seconds). If none, reply [].", "labels": [[6.72, 7.33, "yorkshire"]], "response_mode": "offline", "answer": "[{\"word\":\"yorkshire\",\"start\":6.72,\"end\":7.33}]"} {"utterance_id": "5142-36377-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0015.flac", "task_type": "lexical_detection", "query": "impressions", "prompt": "When the utterance is complete, report all instances of the word \"impressions\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.54, 0.98, "impressions"], [3.52, 4.21, "impressions"]], "response_mode": "offline", "answer": "[{\"word\":\"impressions\",\"start\":0.54,\"end\":0.98},{\"word\":\"impressions\",\"start\":3.52,\"end\":4.21}]"} {"utterance_id": "61-70968-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0032.flac", "task_type": "semantic_detection", "query": "color", "prompt": "You will hear a complete utterance. When it ends, list every instance of a color name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.52, 1.81, "black"]], "response_mode": "offline", "answer": "[{\"word\":\"black\",\"start\":1.52,\"end\":1.81}]"} {"utterance_id": "7729-102255-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0000.flac", "task_type": "lexical_detection", "query": "bogus", "prompt": "Once the audio has finished, identify every instance of the word \"bogus\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.3, 0.69, "bogus"]], "response_mode": "offline", "answer": "[{\"word\":\"bogus\",\"start\":0.3,\"end\":0.69}]"} {"utterance_id": "5639-40744-0039", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0039.flac", "task_type": "lexical_counting", "query": "arms", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for the word \"arms\" with word/start/end. If none, reply [].", "labels": [[4.29, 4.81, "one"], [11.77, 12.22, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.29,\"end\":4.81},{\"word\":\"two\",\"start\":11.77,\"end\":12.22}]"} {"utterance_id": "260-123286-0022", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0022.flac", "task_type": "lexical_detection", "query": "awoke", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"awoke\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.52, 2.87, "awoke"]], "response_mode": "offline", "answer": "[{\"word\":\"awoke\",\"start\":2.52,\"end\":2.87}]"} {"utterance_id": "5105-28241-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28241/5105-28241-0015.flac", "task_type": "lexical_detection", "query": "procope", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"procope\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.82, 3.31, "procope"]], "response_mode": "offline", "answer": "[{\"word\":\"procope\",\"start\":2.82,\"end\":3.31}]"} {"utterance_id": "7176-92135-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0000.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Process the full audio offline. Find every instance of a money or currency term and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.38, 11.79, "charge"], [13.1, 13.43, "price"]], "response_mode": "offline", "answer": "[{\"word\":\"charge\",\"start\":11.38,\"end\":11.79},{\"word\":\"price\",\"start\":13.1,\"end\":13.43}]"} {"utterance_id": "1089-134686-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0023.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Process the full audio offline. Find every instance of a number and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.58, 3.83, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":3.58,\"end\":3.83}]"} {"utterance_id": "2961-961-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0002.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Process the full audio offline. Find every instance of a city name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.0, 4.71, "locris"]], "response_mode": "offline", "answer": "[{\"word\":\"locris\",\"start\":4.0,\"end\":4.71}]"} {"utterance_id": "8455-210777-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0020.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Once the audio has finished, identify every instance of a person's name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.26, 1.77, "jack"]], "response_mode": "offline", "answer": "[{\"word\":\"jack\",\"start\":1.26,\"end\":1.77}]"} {"utterance_id": "4446-2271-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2271/4446-2271-0005.flac", "task_type": "lexical_detection", "query": "act", "prompt": "After listening to the whole recording, output every instance of the word \"act\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.06, 3.32, "act"]], "response_mode": "offline", "answer": "[{\"word\":\"act\",\"start\":3.06,\"end\":3.32}]"} {"utterance_id": "61-70968-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0002.flac", "task_type": "semantic_detection", "query": "money", "prompt": "You will hear a complete utterance. When it ends, list every instance of a money or currency term you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.8, 1.35, "fortune"]], "response_mode": "offline", "answer": "[{\"word\":\"fortune\",\"start\":0.8,\"end\":1.35}]"} {"utterance_id": "1089-134686-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0028.flac", "task_type": "lexical_detection", "query": "wednesday", "prompt": "After the audio ends, extract all occurrences of the word \"wednesday\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.89, 2.47, "wednesday"]], "response_mode": "offline", "answer": "[{\"word\":\"wednesday\",\"start\":1.89,\"end\":2.47}]"} {"utterance_id": "4446-2273-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0009.flac", "task_type": "lexical_detection", "query": "mother's", "prompt": "After listening to the whole recording, output every instance of the word \"mother's\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.32, 3.87, "mother's"]], "response_mode": "offline", "answer": "[{\"word\":\"mother's\",\"start\":3.32,\"end\":3.87}]"} {"utterance_id": "6829-68769-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0041.flac", "task_type": "lexical_detection", "query": "i'm", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"i'm\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.27, 0.46, "i'm"]], "response_mode": "offline", "answer": "[{\"word\":\"i'm\",\"start\":0.27,\"end\":0.46}]"} {"utterance_id": "6930-75918-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6930/75918/6930-75918-0000.flac", "task_type": "lexical_detection", "query": "amidst", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"amidst\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.29, 2.71, "amidst"]], "response_mode": "offline", "answer": "[{\"word\":\"amidst\",\"start\":2.29,\"end\":2.71}]"} {"utterance_id": "1089-134686-0027", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0027.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.09, 1.42, "hands"]], "response_mode": "offline", "answer": "[{\"word\":\"hands\",\"start\":1.09,\"end\":1.42}]"} {"utterance_id": "7127-75947-0030", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0030.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.05, 2.45, "count"]], "response_mode": "offline", "answer": "[{\"word\":\"count\",\"start\":2.05,\"end\":2.45}]"} {"utterance_id": "6829-68769-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0009.flac", "task_type": "lexical_detection", "query": "markham", "prompt": "After listening to the whole recording, output every instance of the word \"markham\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.78, 3.28, "markham"]], "response_mode": "offline", "answer": "[{\"word\":\"markham\",\"start\":2.78,\"end\":3.28}]"} {"utterance_id": "4077-13754-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0001.flac", "task_type": "lexical_detection", "query": "concerning", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"concerning\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.32, 1.89, "concerning"]], "response_mode": "offline", "answer": "[{\"word\":\"concerning\",\"start\":1.32,\"end\":1.89}]"} {"utterance_id": "4970-29095-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0028.flac", "task_type": "semantic_detection", "query": "date", "prompt": "Do not answer until the audio ends. Then list each instance of a date as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[9.8, 10.27, "wednesday"]], "response_mode": "offline", "answer": "[{\"word\":\"wednesday\",\"start\":9.8,\"end\":10.27}]"} {"utterance_id": "2094-142345-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0012.flac", "task_type": "lexical_counting", "query": "between", "prompt": "Listen to the audio in full, then emit a JSON array of count events for the word \"between\" (word, start, end). If none, reply [].", "labels": [[1.58, 1.97, "one"], [5.74, 6.05, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.58,\"end\":1.97},{\"word\":\"two\",\"start\":5.74,\"end\":6.05}]"} {"utterance_id": "7127-75946-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0002.flac", "task_type": "semantic_counting", "query": "profession", "prompt": "After the audio ends, tally every instance of a profession. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[1.94, 2.41, "one"], [10.24, 10.48, "two"], [12.05, 12.78, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.94,\"end\":2.41},{\"word\":\"two\",\"start\":10.24,\"end\":10.48},{\"word\":\"three\",\"start\":12.05,\"end\":12.78}]"} {"utterance_id": "1089-134686-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0000.flac", "task_type": "lexical_detection", "query": "hoped", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"hoped\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.62, 1.01, "hoped"]], "response_mode": "offline", "answer": "[{\"word\":\"hoped\",\"start\":0.62,\"end\":1.01}]"} {"utterance_id": "1089-134686-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1089/134686/1089-134686-0033.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After listening to the whole recording, output every instance of a person's name with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.57, 3.16, "saint francis xavier"]], "response_mode": "offline", "answer": "[{\"word\":\"saint francis xavier\",\"start\":1.57,\"end\":3.16}]"} {"utterance_id": "4446-2273-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2273/4446-2273-0020.flac", "task_type": "lexical_detection", "query": "answered", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"answered\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.6, 1.88, "answered"]], "response_mode": "offline", "answer": "[{\"word\":\"answered\",\"start\":1.6,\"end\":1.88}]"} {"utterance_id": "2300-131720-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0024.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "When the utterance is complete, report all instances of a number including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.67, 2.9, "three"]], "response_mode": "offline", "answer": "[{\"word\":\"three\",\"start\":2.67,\"end\":2.9}]"} {"utterance_id": "3575-170457-0046", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0046.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Once the audio has finished, identify every instance of a body part. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[24.88, 25.24, "leg"]], "response_mode": "offline", "answer": "[{\"word\":\"leg\",\"start\":24.88,\"end\":25.24}]"} {"utterance_id": "5639-40744-0041", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0041.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "After listening to the whole recording, output every instance of a family relation with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[16.14, 16.74, "parents"]], "response_mode": "offline", "answer": "[{\"word\":\"parents\",\"start\":16.14,\"end\":16.74}]"} {"utterance_id": "5105-28233-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5105/28233/5105-28233-0001.flac", "task_type": "lexical_detection", "query": "possessed", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"possessed\" with fields word, start, end (seconds). If none, reply [].", "labels": [[3.56, 4.22, "possessed"]], "response_mode": "offline", "answer": "[{\"word\":\"possessed\",\"start\":3.56,\"end\":4.22}]"} {"utterance_id": "4446-2275-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0002.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a person's name (word, start, end). If none, reply [].", "labels": [[0.35, 0.86, "one"], [6.44, 6.81, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.35,\"end\":0.86},{\"word\":\"two\",\"start\":6.44,\"end\":6.81}]"} {"utterance_id": "61-70968-0060", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0060.flac", "task_type": "lexical_detection", "query": "thanks", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"thanks\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.55, 1.03, "thanks"]], "response_mode": "offline", "answer": "[{\"word\":\"thanks\",\"start\":0.55,\"end\":1.03}]"} {"utterance_id": "1320-122617-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0023.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the full audio. After it finishes, find every instance of a person's name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.45, 1.09, "uncas"]], "response_mode": "offline", "answer": "[{\"word\":\"uncas\",\"start\":0.45,\"end\":1.09}]"} {"utterance_id": "2961-960-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0009.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of an animal name with fields word, start, end (seconds). If none, reply [].", "labels": [[3.3, 3.66, "animal"], [10.64, 11.33, "animals"]], "response_mode": "offline", "answer": "[{\"word\":\"animal\",\"start\":3.3,\"end\":3.66},{\"word\":\"animals\",\"start\":10.64,\"end\":11.33}]"} {"utterance_id": "7176-88083-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0005.flac", "task_type": "lexical_detection", "query": "perch", "prompt": "Once the audio has finished, identify every instance of the word \"perch\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[4.08, 4.53, "perch"]], "response_mode": "offline", "answer": "[{\"word\":\"perch\",\"start\":4.08,\"end\":4.53}]"} {"utterance_id": "6829-68771-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0002.flac", "task_type": "lexical_detection", "query": "strengthened", "prompt": "When the utterance is complete, report all instances of the word \"strengthened\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.28, 2.87, "strengthened"]], "response_mode": "offline", "answer": "[{\"word\":\"strengthened\",\"start\":2.28,\"end\":2.87}]"} {"utterance_id": "4970-29095-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0038.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Do not answer until the audio ends. Then return a JSON list of count events for a person's name with word/start/end. If none, reply [].", "labels": [[0.53, 0.84, "one"], [2.14, 2.45, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.53,\"end\":0.84},{\"word\":\"two\",\"start\":2.14,\"end\":2.45}]"} {"utterance_id": "4446-2275-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0009.flac", "task_type": "lexical_detection", "query": "ago", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"ago\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.62, 2.0, "ago"]], "response_mode": "offline", "answer": "[{\"word\":\"ago\",\"start\":1.62,\"end\":2.0}]"} {"utterance_id": "7127-75947-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75947/7127-75947-0023.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.11, 3.63, "mademoiselle de tonnay charente"]], "response_mode": "offline", "answer": "[{\"word\":\"mademoiselle de tonnay charente\",\"start\":2.11,\"end\":3.63}]"} {"utterance_id": "7127-75946-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0002.flac", "task_type": "lexical_detection", "query": "courtiers", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"courtiers\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[12.05, 12.78, "courtiers"]], "response_mode": "offline", "answer": "[{\"word\":\"courtiers\",\"start\":12.05,\"end\":12.78}]"} {"utterance_id": "61-70968-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70968/61-70968-0023.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "Listen to the full audio. After it finishes, find every instance of an ordinal number. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.23, 4.54, "second"]], "response_mode": "offline", "answer": "[{\"word\":\"second\",\"start\":4.23,\"end\":4.54}]"} {"utterance_id": "7729-102255-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0012.flac", "task_type": "semantic_detection", "query": "state_or_province", "prompt": "When the utterance is complete, report all instances of a state or province name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.27, 1.6, "state"]], "response_mode": "offline", "answer": "[{\"word\":\"state\",\"start\":1.27,\"end\":1.6}]"} {"utterance_id": "8555-292519-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0015.flac", "task_type": "lexical_detection", "query": "broken", "prompt": "After listening to the whole recording, output every instance of the word \"broken\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.76, 1.2, "broken"]], "response_mode": "offline", "answer": "[{\"word\":\"broken\",\"start\":0.76,\"end\":1.2}]"} {"utterance_id": "8555-284449-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0020.flac", "task_type": "semantic_detection", "query": "food", "prompt": "Listen to the full audio. After it finishes, find every instance of a food name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.69, 4.17, "supper"]], "response_mode": "offline", "answer": "[{\"word\":\"supper\",\"start\":3.69,\"end\":4.17}]"} {"utterance_id": "672-122797-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0034.flac", "task_type": "lexical_detection", "query": "drawing", "prompt": "When the utterance is complete, report all instances of the word \"drawing\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.2, 2.58, "drawing"]], "response_mode": "offline", "answer": "[{\"word\":\"drawing\",\"start\":2.2,\"end\":2.58}]"} {"utterance_id": "8455-210777-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0023.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "After the audio ends, extract all occurrences of a profession. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.06, 1.61, "officers"]], "response_mode": "offline", "answer": "[{\"word\":\"officers\",\"start\":1.06,\"end\":1.61}]"} {"utterance_id": "8230-279154-0037", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0037.flac", "task_type": "lexical_counting", "query": "cat", "prompt": "Listen to the entire clip, then report the count sequence for the word \"cat\" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.53, 12.09, "one"], [18.81, 19.28, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":11.53,\"end\":12.09},{\"word\":\"two\",\"start\":18.81,\"end\":19.28}]"} {"utterance_id": "237-134500-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134500/237-134500-0012.flac", "task_type": "lexical_counting", "query": "few", "prompt": "You will hear a complete utterance. When it ends, produce a running count of the word \"few\": each event is {\"word\": \"one\"|\"two\"|..., \"start\", \"end\"}. If none, reply [].", "labels": [[0.5, 0.73, "one"], [8.63, 8.91, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.5,\"end\":0.73},{\"word\":\"two\",\"start\":8.63,\"end\":8.91}]"} {"utterance_id": "7176-88083-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0004.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Process the full audio offline. Find every instance of an animal name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.14, 3.66, "cormorant"]], "response_mode": "offline", "answer": "[{\"word\":\"cormorant\",\"start\":3.14,\"end\":3.66}]"} {"utterance_id": "2094-142345-0010", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0010.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "You will hear a complete utterance. When it ends, list every instance of a plant name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[11.2, 11.47, "oak"]], "response_mode": "offline", "answer": "[{\"word\":\"oak\",\"start\":11.2,\"end\":11.47}]"} {"utterance_id": "908-31957-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0012.flac", "task_type": "lexical_detection", "query": "star", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"star\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[5.38, 6.08, "star"]], "response_mode": "offline", "answer": "[{\"word\":\"star\",\"start\":5.38,\"end\":6.08}]"} {"utterance_id": "7176-92135-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0001.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Do not answer until the audio ends. Then list each instance of a city name as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.78, 3.07, "london"]], "response_mode": "offline", "answer": "[{\"word\":\"london\",\"start\":2.78,\"end\":3.07}]"} {"utterance_id": "3729-6852-0032", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0032.flac", "task_type": "lexical_detection", "query": "shop", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"shop\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.09, 4.51, "shop"]], "response_mode": "offline", "answer": "[{\"word\":\"shop\",\"start\":4.09,\"end\":4.51}]"} {"utterance_id": "7176-88083-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/88083/7176-88083-0008.flac", "task_type": "lexical_detection", "query": "downward", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"downward\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.87, 2.36, "downward"]], "response_mode": "offline", "answer": "[{\"word\":\"downward\",\"start\":1.87,\"end\":2.36}]"} {"utterance_id": "6829-68769-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0003.flac", "task_type": "lexical_detection", "query": "deliberate", "prompt": "When the utterance is complete, report all instances of the word \"deliberate\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.53, 1.11, "deliberate"]], "response_mode": "offline", "answer": "[{\"word\":\"deliberate\",\"start\":0.53,\"end\":1.11}]"} {"utterance_id": "6829-68769-0034", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0034.flac", "task_type": "lexical_detection", "query": "assist", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"assist\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.78, 4.19, "assist"]], "response_mode": "offline", "answer": "[{\"word\":\"assist\",\"start\":3.78,\"end\":4.19}]"} {"utterance_id": "4507-16021-0036", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4507/16021/4507-16021-0036.flac", "task_type": "lexical_detection", "query": "facts", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"facts\" with fields word, start, end (seconds). If none, reply [].", "labels": [[0.31, 0.93, "facts"]], "response_mode": "offline", "answer": "[{\"word\":\"facts\",\"start\":0.31,\"end\":0.93}]"} {"utterance_id": "4992-41797-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41797/4992-41797-0017.flac", "task_type": "lexical_detection", "query": "worry", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"worry\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.43, 3.01, "worry"]], "response_mode": "offline", "answer": "[{\"word\":\"worry\",\"start\":2.43,\"end\":3.01}]"} {"utterance_id": "121-121726-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/121/121726/121-121726-0011.flac", "task_type": "semantic_counting", "query": "kinship", "prompt": "After the audio ends, tally every instance of a family relation. Reply with JSON only — one object per occurrence, word=\"one\"/\"two\"/... plus start and end in seconds. If none, reply [].", "labels": [[0.51, 1.37, "one"], [3.3, 3.83, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.51,\"end\":1.37},{\"word\":\"two\",\"start\":3.3,\"end\":3.83}]"} {"utterance_id": "1188-133604-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0029.flac", "task_type": "lexical_detection", "query": "sorts", "prompt": "After the audio ends, extract all occurrences of the word \"sorts\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.91, 2.22, "sorts"]], "response_mode": "offline", "answer": "[{\"word\":\"sorts\",\"start\":1.91,\"end\":2.22}]"} {"utterance_id": "237-134493-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/134493/237-134493-0007.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen carefully to the entire clip, then report every instance of a person's name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.18, 0.83, "alexandra"]], "response_mode": "offline", "answer": "[{\"word\":\"alexandra\",\"start\":0.18,\"end\":0.83}]"} {"utterance_id": "7127-75946-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0017.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "Listen to the full audio. After it finishes, find every instance of an organization name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.42, 2.89, "court"]], "response_mode": "offline", "answer": "[{\"word\":\"court\",\"start\":2.42,\"end\":2.89}]"} {"utterance_id": "1995-1837-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0028.flac", "task_type": "lexical_detection", "query": "empty", "prompt": "When the utterance is complete, report all instances of the word \"empty\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.78, 1.18, "empty"]], "response_mode": "offline", "answer": "[{\"word\":\"empty\",\"start\":0.78,\"end\":1.18}]"} {"utterance_id": "1995-1837-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0007.flac", "task_type": "lexical_detection", "query": "shot", "prompt": "Once the audio has finished, identify every instance of the word \"shot\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[2.32, 2.67, "shot"]], "response_mode": "offline", "answer": "[{\"word\":\"shot\",\"start\":2.32,\"end\":2.67}]"} {"utterance_id": "8555-284447-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284447/8555-284447-0014.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a person's name with fields word, start, end (seconds). If none, reply [].", "labels": [[1.55, 2.06, "cap'n bill"]], "response_mode": "offline", "answer": "[{\"word\":\"cap'n bill\",\"start\":1.55,\"end\":2.06}]"} {"utterance_id": "61-70970-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0007.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After the audio ends, extract all occurrences of a body part. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.92, 4.29, "arm"]], "response_mode": "offline", "answer": "[{\"word\":\"arm\",\"start\":3.92,\"end\":4.29}]"} {"utterance_id": "4077-13751-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13751/4077-13751-0015.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "When the utterance is complete, report all instances of an organization name including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[8.31, 8.74, "company"]], "response_mode": "offline", "answer": "[{\"word\":\"company\",\"start\":8.31,\"end\":8.74}]"} {"utterance_id": "3570-5696-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3570/5696/3570-5696-0005.flac", "task_type": "lexical_detection", "query": "entire", "prompt": "Process the full audio offline. Find every instance of the word \"entire\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.85, 1.31, "entire"]], "response_mode": "offline", "answer": "[{\"word\":\"entire\",\"start\":0.85,\"end\":1.31}]"} {"utterance_id": "2961-960-0017", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0017.flac", "task_type": "lexical_detection", "query": "accustomed", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"accustomed\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.49, 7.12, "accustomed"]], "response_mode": "offline", "answer": "[{\"word\":\"accustomed\",\"start\":6.49,\"end\":7.12}]"} {"utterance_id": "4446-2275-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0008.flac", "task_type": "lexical_detection", "query": "bartley", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"bartley\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.83, 1.24, "bartley"]], "response_mode": "offline", "answer": "[{\"word\":\"bartley\",\"start\":0.83,\"end\":1.24}]"} {"utterance_id": "5142-33396-0068", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0068.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "When the utterance is complete, report all instances of a body part including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.64, 1.97, "tooth"]], "response_mode": "offline", "answer": "[{\"word\":\"tooth\",\"start\":1.64,\"end\":1.97}]"} {"utterance_id": "1320-122617-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122617/1320-122617-0002.flac", "task_type": "semantic_counting", "query": "animal", "prompt": "Listen to the full audio. Count every occurrence of an animal name. After it finishes, for each occurrence in order output a JSON object whose word is the count word (\"one\", \"two\", ...) with that occurrence's start/end times. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.99, 4.32, "one"], [7.34, 7.71, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":3.99,\"end\":4.32},{\"word\":\"two\",\"start\":7.34,\"end\":7.71}]"} {"utterance_id": "2300-131720-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2300/131720/2300-131720-0023.flac", "task_type": "semantic_detection", "query": "clock_time", "prompt": "Do not answer until the audio ends. Then list each instance of a time of day as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[8.54, 8.94, "morning"]], "response_mode": "offline", "answer": "[{\"word\":\"morning\",\"start\":8.54,\"end\":8.94}]"} {"utterance_id": "8463-294828-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294828/8463-294828-0018.flac", "task_type": "lexical_detection", "query": "we'll", "prompt": "Once the audio has finished, identify every instance of the word \"we'll\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.34, 0.54, "we'll"]], "response_mode": "offline", "answer": "[{\"word\":\"we'll\",\"start\":0.34,\"end\":0.54}]"} {"utterance_id": "3575-170457-0053", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0053.flac", "task_type": "semantic_detection", "query": "animal", "prompt": "Listen to the full audio. After it finishes, find every instance of an animal name. For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[7.71, 8.22, "creature"]], "response_mode": "offline", "answer": "[{\"word\":\"creature\",\"start\":7.71,\"end\":8.22}]"} {"utterance_id": "1580-141083-0051", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0051.flac", "task_type": "lexical_detection", "query": "minute", "prompt": "Once the audio has finished, identify every instance of the word \"minute\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.92, 1.23, "minute"]], "response_mode": "offline", "answer": "[{\"word\":\"minute\",\"start\":0.92,\"end\":1.23}]"} {"utterance_id": "1320-122612-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0016.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.87, 3.31, "foot"]], "response_mode": "offline", "answer": "[{\"word\":\"foot\",\"start\":2.87,\"end\":3.31}]"} {"utterance_id": "7127-75946-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7127/75946/7127-75946-0004.flac", "task_type": "lexical_detection", "query": "sire", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"sire\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.94, 1.37, "sire"]], "response_mode": "offline", "answer": "[{\"word\":\"sire\",\"start\":0.94,\"end\":1.37}]"} {"utterance_id": "908-31957-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/31957/908-31957-0003.flac", "task_type": "lexical_detection", "query": "hastily", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"hastily\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.43, 3.09, "hastily"]], "response_mode": "offline", "answer": "[{\"word\":\"hastily\",\"start\":2.43,\"end\":3.09}]"} {"utterance_id": "1188-133604-0025", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0025.flac", "task_type": "lexical_detection", "query": "telling", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"telling\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.72, 2.04, "telling"]], "response_mode": "offline", "answer": "[{\"word\":\"telling\",\"start\":1.72,\"end\":2.04}]"} {"utterance_id": "1221-135767-0005", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1221/135767/1221-135767-0005.flac", "task_type": "lexical_counting", "query": "letter", "prompt": "After listening to the whole recording, output timed count words for the word \"letter\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[1.27, 1.67, "one"], [3.74, 4.04, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":1.27,\"end\":1.67},{\"word\":\"two\",\"start\":3.74,\"end\":4.04}]"} {"utterance_id": "7176-92135-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7176/92135/7176-92135-0000.flac", "task_type": "lexical_detection", "query": "elect", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"elect\" with fields word, start, end (seconds). If none, reply [].", "labels": [[2.51, 2.89, "elect"]], "response_mode": "offline", "answer": "[{\"word\":\"elect\",\"start\":2.51,\"end\":2.89}]"} {"utterance_id": "6829-68769-0024", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68769/6829-68769-0024.flac", "task_type": "semantic_counting", "query": "person_name", "prompt": "Listen to the entire clip, then report the count sequence for a person's name as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.42, 0.65, "one"], [3.0, 3.7, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.42,\"end\":0.65},{\"word\":\"two\",\"start\":3.0,\"end\":3.7}]"} {"utterance_id": "1284-1180-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1180/1284-1180-0008.flac", "task_type": "semantic_counting", "query": "clock_time", "prompt": "When the utterance is complete, report how a time of day accumulated over time using count words and spans. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.73, 1.14, "one"], [3.15, 3.57, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.73,\"end\":1.14},{\"word\":\"two\",\"start\":3.15,\"end\":3.57}]"} {"utterance_id": "1580-141083-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0006.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[6.93, 7.25, "ten"]], "response_mode": "offline", "answer": "[{\"word\":\"ten\",\"start\":6.93,\"end\":7.25}]"} {"utterance_id": "1580-141083-0044", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1580/141083/1580-141083-0044.flac", "task_type": "lexical_detection", "query": "least", "prompt": "Listen to the full audio. After it finishes, find every instance of the word \"least\". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.04, 4.37, "least"]], "response_mode": "offline", "answer": "[{\"word\":\"least\",\"start\":4.04,\"end\":4.37}]"} {"utterance_id": "5142-36377-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0023.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "Process the full audio offline. Find every instance of a person's name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.97, 2.36, "ambrose"], [3.6, 4.38, "john jago"]], "response_mode": "offline", "answer": "[{\"word\":\"ambrose\",\"start\":1.97,\"end\":2.36},{\"word\":\"john jago\",\"start\":3.6,\"end\":4.38}]"} {"utterance_id": "237-126133-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0020.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "When the utterance is complete, report all instances of a family relation including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.64, 1.0, "mother"]], "response_mode": "offline", "answer": "[{\"word\":\"mother\",\"start\":0.64,\"end\":1.0}]"} {"utterance_id": "2830-3980-0012", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2830/3980/2830-3980-0012.flac", "task_type": "lexical_detection", "query": "claim", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"claim\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[1.13, 1.65, "claim"]], "response_mode": "offline", "answer": "[{\"word\":\"claim\",\"start\":1.13,\"end\":1.65}]"} {"utterance_id": "1320-122612-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0006.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "You will hear a complete utterance. When it ends, list every instance of a body part you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.12, 4.68, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"eyes\",\"start\":4.12,\"end\":4.68}]"} {"utterance_id": "3575-170457-0016", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3575/170457/3575-170457-0016.flac", "task_type": "lexical_detection", "query": "farewell", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"farewell\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.52, 1.04, "farewell"]], "response_mode": "offline", "answer": "[{\"word\":\"farewell\",\"start\":0.52,\"end\":1.04}]"} {"utterance_id": "2961-961-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/961/2961-961-0014.flac", "task_type": "lexical_detection", "query": "ours", "prompt": "You will hear a complete utterance. When it ends, list every instance of the word \"ours\" you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.98, 7.61, "ours"]], "response_mode": "offline", "answer": "[{\"word\":\"ours\",\"start\":6.98,\"end\":7.61}]"} {"utterance_id": "1995-1837-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1995/1837/1995-1837-0000.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Process the full audio offline. Find every instance of a color name and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.64, 1.04, "silver"]], "response_mode": "offline", "answer": "[{\"word\":\"silver\",\"start\":0.64,\"end\":1.04}]"} {"utterance_id": "1188-133604-0015", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1188/133604/1188-133604-0015.flac", "task_type": "lexical_detection", "query": "painting", "prompt": "Process the full audio offline. Find every instance of the word \"painting\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[13.43, 13.82, "painting"]], "response_mode": "offline", "answer": "[{\"word\":\"painting\",\"start\":13.43,\"end\":13.82}]"} {"utterance_id": "260-123286-0001", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/260/123286/260-123286-0001.flac", "task_type": "lexical_detection", "query": "extremely", "prompt": "After listening to the whole recording, output every instance of the word \"extremely\" with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.32, 1.96, "extremely"]], "response_mode": "offline", "answer": "[{\"word\":\"extremely\",\"start\":1.32,\"end\":1.96}]"} {"utterance_id": "237-126133-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0021.flac", "task_type": "lexical_detection", "query": "didn't", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"didn't\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.92, 2.15, "didn't"]], "response_mode": "offline", "answer": "[{\"word\":\"didn't\",\"start\":1.92,\"end\":2.15}]"} {"utterance_id": "5683-32865-0006", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32865/5683-32865-0006.flac", "task_type": "lexical_detection", "query": "amusing", "prompt": "Process the full audio offline. Find every instance of the word \"amusing\" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.38, 3.1, "amusing"]], "response_mode": "offline", "answer": "[{\"word\":\"amusing\",\"start\":2.38,\"end\":3.1}]"} {"utterance_id": "672-122797-0007", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0007.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "Once the audio has finished, identify every instance of a plant name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.41, 5.9, "trees"]], "response_mode": "offline", "answer": "[{\"word\":\"trees\",\"start\":5.41,\"end\":5.9}]"} {"utterance_id": "7021-85628-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7021/85628/7021-85628-0023.flac", "task_type": "semantic_detection", "query": "color", "prompt": "After the audio ends, extract all occurrences of a color name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.64, 5.05, "golden"]], "response_mode": "offline", "answer": "[{\"word\":\"golden\",\"start\":4.64,\"end\":5.05}]"} {"utterance_id": "8230-279154-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8230/279154/8230-279154-0018.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[10.06, 10.34, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":10.06,\"end\":10.34}]"} {"utterance_id": "4992-41806-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4992/41806/4992-41806-0013.flac", "task_type": "lexical_detection", "query": "approaching", "prompt": "Once the audio has finished, identify every instance of the word \"approaching\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.53, 1.15, "approaching"]], "response_mode": "offline", "answer": "[{\"word\":\"approaching\",\"start\":0.53,\"end\":1.15}]"} {"utterance_id": "61-70970-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0020.flac", "task_type": "lexical_detection", "query": "opening", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"opening\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[2.07, 2.45, "opening"]], "response_mode": "offline", "answer": "[{\"word\":\"opening\",\"start\":2.07,\"end\":2.45}]"} {"utterance_id": "3729-6852-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0028.flac", "task_type": "lexical_detection", "query": "honest", "prompt": "Once the audio has finished, identify every instance of the word \"honest\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[0.8, 1.19, "honest"]], "response_mode": "offline", "answer": "[{\"word\":\"honest\",\"start\":0.8,\"end\":1.19}]"} {"utterance_id": "5142-36377-0018", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/36377/5142-36377-0018.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "After the audio ends, extract all occurrences of a person's name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.84, 1.26, "naomi"]], "response_mode": "offline", "answer": "[{\"word\":\"naomi\",\"start\":0.84,\"end\":1.26}]"} {"utterance_id": "8455-210777-0065", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8455/210777/8455-210777-0065.flac", "task_type": "lexical_detection", "query": "charge", "prompt": "When the utterance is complete, report all instances of the word \"charge\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.67, 1.94, "charge"]], "response_mode": "offline", "answer": "[{\"word\":\"charge\",\"start\":1.67,\"end\":1.94}]"} {"utterance_id": "3729-6852-0009", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/3729/6852/3729-6852-0009.flac", "task_type": "lexical_detection", "query": "miracle", "prompt": "When the utterance is complete, report all instances of the word \"miracle\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[9.88, 10.5, "miracle"]], "response_mode": "offline", "answer": "[{\"word\":\"miracle\",\"start\":9.88,\"end\":10.5}]"} {"utterance_id": "4077-13754-0013", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0013.flac", "task_type": "semantic_detection", "query": "ordinal", "prompt": "After the audio ends, extract all occurrences of an ordinal number. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[4.84, 5.16, "last"]], "response_mode": "offline", "answer": "[{\"word\":\"last\",\"start\":4.84,\"end\":5.16}]"} {"utterance_id": "4446-2275-0043", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0043.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "Process the full audio offline. Find every instance of a body part and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.12, 2.52, "arms"], [3.54, 3.92, "mouth"], [5.0, 5.47, "eyes"]], "response_mode": "offline", "answer": "[{\"word\":\"arms\",\"start\":2.12,\"end\":2.52},{\"word\":\"mouth\",\"start\":3.54,\"end\":3.92},{\"word\":\"eyes\",\"start\":5.0,\"end\":5.47}]"} {"utterance_id": "4446-2275-0028", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4446/2275/4446-2275-0028.flac", "task_type": "semantic_detection", "query": "body_part", "prompt": "After listening to the whole recording, output every instance of a body part with precise start and end times (seconds). Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.19, 2.39, "hand"]], "response_mode": "offline", "answer": "[{\"word\":\"hand\",\"start\":2.19,\"end\":2.39}]"} {"utterance_id": "2961-960-0011", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2961/960/2961-960-0011.flac", "task_type": "lexical_counting", "query": "like", "prompt": "After listening to the whole recording, output timed count words for the word \"like\" (one, two, three, ...) as JSON. If none, reply [].", "labels": [[4.68, 5.01, "one"], [14.99, 15.27, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":4.68,\"end\":5.01},{\"word\":\"two\",\"start\":14.99,\"end\":15.27}]"} {"utterance_id": "4970-29095-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0026.flac", "task_type": "semantic_counting", "query": "plant", "prompt": "Listen to the audio in full, then emit a JSON array of count events for a plant name (word, start, end). If none, reply [].", "labels": [[7.99, 8.58, "one"], [10.12, 10.66, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":7.99,\"end\":8.58},{\"word\":\"two\",\"start\":10.12,\"end\":10.66}]"} {"utterance_id": "237-126133-0004", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0004.flac", "task_type": "lexical_detection", "query": "phronsie", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of the word \"phronsie\" with fields word, start, end (seconds). If none, reply [].", "labels": [[1.21, 1.69, "phronsie"]], "response_mode": "offline", "answer": "[{\"word\":\"phronsie\",\"start\":1.21,\"end\":1.69}]"} {"utterance_id": "1320-122612-0002", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0002.flac", "task_type": "semantic_detection", "query": "person_name", "prompt": "You will hear a complete utterance. When it ends, list every instance of a person's name you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.01, 3.56, "hawkeye"]], "response_mode": "offline", "answer": "[{\"word\":\"hawkeye\",\"start\":3.01,\"end\":3.56}]"} {"utterance_id": "1284-1181-0008", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1284/1181/1284-1181-0008.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Once the audio has finished, identify every instance of a profession. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.05, 5.7, "servant"]], "response_mode": "offline", "answer": "[{\"word\":\"servant\",\"start\":5.05,\"end\":5.7}]"} {"utterance_id": "5639-40744-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5639/40744/5639-40744-0019.flac", "task_type": "lexical_detection", "query": "judgment", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"judgment\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[11.13, 11.73, "judgment"]], "response_mode": "offline", "answer": "[{\"word\":\"judgment\",\"start\":11.13,\"end\":11.73}]"} {"utterance_id": "7729-102255-0038", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/7729/102255/7729-102255-0038.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "Do not answer until the audio ends. Then list each instance of a number as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[4.21, 4.41, "two"]], "response_mode": "offline", "answer": "[{\"word\":\"two\",\"start\":4.21,\"end\":4.41}]"} {"utterance_id": "8463-294825-0014", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0014.flac", "task_type": "lexical_detection", "query": "six", "prompt": "After the audio ends, extract all occurrences of the word \"six\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.1, 1.51, "six"]], "response_mode": "offline", "answer": "[{\"word\":\"six\",\"start\":1.1,\"end\":1.51}]"} {"utterance_id": "8555-284449-0020", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/284449/8555-284449-0020.flac", "task_type": "semantic_detection", "query": "organization", "prompt": "After the audio ends, extract all occurrences of an organization name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.91, 1.31, "bands"]], "response_mode": "offline", "answer": "[{\"word\":\"bands\",\"start\":0.91,\"end\":1.31}]"} {"utterance_id": "5683-32866-0029", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5683/32866/5683-32866-0029.flac", "task_type": "lexical_detection", "query": "nervous", "prompt": "When the utterance is complete, report all instances of the word \"nervous\" including the spoken word and time span. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[1.36, 2.0, "nervous"]], "response_mode": "offline", "answer": "[{\"word\":\"nervous\",\"start\":1.36,\"end\":2.0}]"} {"utterance_id": "4970-29095-0003", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4970/29095/4970-29095-0003.flac", "task_type": "semantic_detection", "query": "profession", "prompt": "Listen to the audio in full, then emit a JSON array of every instance of a profession with fields word, start, end (seconds). If none, reply [].", "labels": [[1.63, 2.03, "elders"]], "response_mode": "offline", "answer": "[{\"word\":\"elders\",\"start\":1.63,\"end\":2.03}]"} {"utterance_id": "908-157963-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/908/157963/908-157963-0026.flac", "task_type": "lexical_detection", "query": "loved", "prompt": "Once the audio has finished, identify every instance of the word \"loved\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[3.88, 4.21, "loved"]], "response_mode": "offline", "answer": "[{\"word\":\"loved\",\"start\":3.88,\"end\":4.21}]"} {"utterance_id": "237-126133-0023", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/237/126133/237-126133-0023.flac", "task_type": "semantic_detection", "query": "plant", "prompt": "After the audio ends, extract all occurrences of a plant name. Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.33, 3.86, "peppers"]], "response_mode": "offline", "answer": "[{\"word\":\"peppers\",\"start\":3.33,\"end\":3.86}]"} {"utterance_id": "2094-142345-0050", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/2094/142345/2094-142345-0050.flac", "task_type": "semantic_detection", "query": "kinship", "prompt": "Process the full audio offline. Find every instance of a family relation and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.8, 1.23, "father"]], "response_mode": "offline", "answer": "[{\"word\":\"father\",\"start\":0.8,\"end\":1.23}]"} {"utterance_id": "8555-292519-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8555/292519/8555-292519-0000.flac", "task_type": "semantic_detection", "query": "color", "prompt": "Once the audio has finished, identify every instance of a color name. Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[5.04, 5.54, "color"]], "response_mode": "offline", "answer": "[{\"word\":\"color\",\"start\":5.04,\"end\":5.54}]"} {"utterance_id": "672-122797-0066", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/672/122797/672-122797-0066.flac", "task_type": "semantic_detection", "query": "cardinal", "prompt": "You will hear a complete utterance. When it ends, list every instance of a number you heard with word-level timestamps. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[0.7, 0.99, "one"]], "response_mode": "offline", "answer": "[{\"word\":\"one\",\"start\":0.7,\"end\":0.99}]"} {"utterance_id": "4077-13754-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/4077/13754/4077-13754-0000.flac", "task_type": "semantic_detection", "query": "money", "prompt": "Listen carefully to the entire clip, then report every instance of a money or currency term as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[3.98, 4.42, "wealth"]], "response_mode": "offline", "answer": "[{\"word\":\"wealth\",\"start\":3.98,\"end\":4.42}]"} {"utterance_id": "8463-294825-0019", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/8463/294825/8463-294825-0019.flac", "task_type": "lexical_detection", "query": "inch", "prompt": "Listen carefully to the entire clip, then report every instance of the word \"inch\" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.76, 3.22, "inch"]], "response_mode": "offline", "answer": "[{\"word\":\"inch\",\"start\":2.76,\"end\":3.22}]"} {"utterance_id": "5142-33396-0026", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/5142/33396/5142-33396-0026.flac", "task_type": "lexical_detection", "query": "stingy", "prompt": "After the audio ends, extract all occurrences of the word \"stingy\". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[2.01, 2.56, "stingy"]], "response_mode": "offline", "answer": "[{\"word\":\"stingy\",\"start\":2.01,\"end\":2.56}]"} {"utterance_id": "1320-122612-0000", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/1320/122612/1320-122612-0000.flac", "task_type": "semantic_detection", "query": "city", "prompt": "Listen carefully to the entire clip, then report every instance of a city name as JSON objects with word, start, and end. Reply with JSON only: a list of objects {\"word\": , \"start\": , \"end\": } in chronological order. If none, reply [].", "labels": [[6.46, 7.2, "settlements"]], "response_mode": "offline", "answer": "[{\"word\":\"settlements\",\"start\":6.46,\"end\":7.2}]"} {"utterance_id": "61-70970-0033", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/61/70970/61-70970-0033.flac", "task_type": "lexical_detection", "query": "worth", "prompt": "Once the audio has finished, identify every instance of the word \"worth\". Return a JSON list of {\"word\", \"start\", \"end\"} in time order. If none, reply [].", "labels": [[1.57, 1.93, "worth"]], "response_mode": "offline", "answer": "[{\"word\":\"worth\",\"start\":1.57,\"end\":1.93}]"} {"utterance_id": "6829-68771-0021", "split": "test-clean", "audio_path": "LibriSpeech/test-clean/6829/68771/6829-68771-0021.flac", "task_type": "lexical_detection", "query": "can't", "prompt": "Do not answer until the audio ends. Then list each instance of the word \"can't\" as {\"word\", \"start\", \"end\"}. If none, reply [].", "labels": [[0.86, 1.17, "can't"]], "response_mode": "offline", "answer": "[{\"word\":\"can't\",\"start\":0.86,\"end\":1.17}]"}