utterance_id stringlengths 14 16 | split stringclasses 1
value | audio_path stringlengths 51 55 | task_type stringclasses 4
values | query stringlengths 3 17 | prompt stringlengths 117 375 | labels listlengths 1 5 | response_mode stringclasses 1
value | answer stringlengths 38 237 |
|---|---|---|---|---|---|---|---|---|
84-121123-0002 | dev-clean | LibriSpeech/dev-clean/84/121123/84-121123-0002.flac | semantic_detection | disease | 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 []. | [
[
9.78,
10.39,
"epilepsy"
]
] | offline | [{"word":"epilepsy","start":9.78,"end":10.39}] |
5895-34615-0008 | dev-clean | LibriSpeech/dev-clean/5895/34615/5895-34615-0008.flac | lexical_detection | depend | Listen to the full audio. After it finishes, find every instance of the word "depend". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.4,
1.88,
"depend"
]
] | offline | [{"word":"depend","start":1.4,"end":1.88}] |
3081-166546-0006 | dev-clean | LibriSpeech/dev-clean/3081/166546/3081-166546-0006.flac | semantic_detection | kinship | 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 []. | [
[
2.5,
2.85,
"father"
]
] | offline | [{"word":"father","start":2.5,"end":2.85}] |
652-129742-0007 | dev-clean | LibriSpeech/dev-clean/652/129742/652-129742-0007.flac | lexical_detection | garnish | Once the audio has finished, identify every instance of the word "garnish". Return a JSON list of {"word", "start", "end"} in time order. If none, reply []. | [
[
0.99,
1.46,
"garnish"
]
] | offline | [{"word":"garnish","start":0.99,"end":1.46}] |
2086-149220-0020 | dev-clean | LibriSpeech/dev-clean/2086/149220/2086-149220-0020.flac | lexical_detection | turn | Listen to the audio in full, then emit a JSON array of every instance of the word "turn" with fields word, start, end (seconds). If none, reply []. | [
[
0.61,
0.85,
"turn"
]
] | offline | [{"word":"turn","start":0.61,"end":0.85}] |
2803-161169-0008 | dev-clean | LibriSpeech/dev-clean/2803/161169/2803-161169-0008.flac | lexical_counting | nut | Listen to the entire clip, then report the count sequence for the word "nut" as timed JSON events (word is the count word, not the surface form). Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
3.75,
4.14,
"one"
],
[
9.6,
10.06,
"two"
]
] | offline | [{"word":"one","start":3.75,"end":4.14},{"word":"two","start":9.6,"end":10.06}] |
5694-64029-0021 | dev-clean | LibriSpeech/dev-clean/5694/64029/5694-64029-0021.flac | semantic_detection | state_or_province | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
2.06,
2.47,
"tennessee"
]
] | offline | [{"word":"tennessee","start":2.06,"end":2.47}] |
5694-64029-0024 | dev-clean | LibriSpeech/dev-clean/5694/64029/5694-64029-0024.flac | lexical_detection | glory | When the utterance is complete, report all instances of the word "glory" including the spoken word and time span. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
2.3,
2.87,
"glory"
]
] | offline | [{"word":"glory","start":2.3,"end":2.87}] |
6295-244435-0014 | dev-clean | LibriSpeech/dev-clean/6295/244435/6295-244435-0014.flac | semantic_detection | person_name | 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 []. | [
[
0.22,
0.76,
"beauregard"
]
] | offline | [{"word":"beauregard","start":0.22,"end":0.76}] |
1272-135031-0003 | dev-clean | LibriSpeech/dev-clean/1272/135031/1272-135031-0003.flac | lexical_detection | asleep | You will hear a complete utterance. When it ends, list every instance of the word "asleep" you heard with word-level timestamps. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.26,
1.83,
"asleep"
]
] | offline | [{"word":"asleep","start":1.26,"end":1.83}] |
2277-149896-0026 | dev-clean | LibriSpeech/dev-clean/2277/149896/2277-149896-0026.flac | semantic_detection | weather | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
0.58,
0.93,
"drizzle"
]
] | offline | [{"word":"drizzle","start":0.58,"end":0.93}] |
2902-9006-0006 | dev-clean | LibriSpeech/dev-clean/2902/9006/2902-9006-0006.flac | semantic_detection | profession | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
0.9400000000000001,
1.46,
"emperors"
]
] | offline | [{"word":"emperors","start":0.94,"end":1.46}] |
6313-66125-0025 | dev-clean | LibriSpeech/dev-clean/6313/66125/6313-66125-0025.flac | semantic_detection | body_part | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
3.02,
3.47,
"hands"
]
] | offline | [{"word":"hands","start":3.02,"end":3.47}] |
5338-24640-0007 | dev-clean | LibriSpeech/dev-clean/5338/24640/5338-24640-0007.flac | semantic_detection | organization | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
16.62,
17.37,
"inquisition"
]
] | offline | [{"word":"inquisition","start":16.62,"end":17.37}] |
2412-153947-0010 | dev-clean | LibriSpeech/dev-clean/2412/153947/2412-153947-0010.flac | semantic_detection | date | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
4.44,
6.24,
"may first eighteen seventy one"
],
[
15.69,
16.72,
"eighteen seventy two"
]
] | offline | [{"word":"may first eighteen seventy one","start":4.44,"end":6.24},{"word":"eighteen seventy two","start":15.69,"end":16.72}] |
3576-138058-0010 | dev-clean | LibriSpeech/dev-clean/3576/138058/3576-138058-0010.flac | lexical_detection | removed | Once the audio has finished, identify every instance of the word "removed". Return a JSON list of {"word", "start", "end"} in time order. If none, reply []. | [
[
1.6400000000000001,
2.24,
"removed"
]
] | offline | [{"word":"removed","start":1.64,"end":2.24}] |
5536-43363-0018 | dev-clean | LibriSpeech/dev-clean/5536/43363/5536-43363-0018.flac | lexical_detection | former | After listening to the whole recording, output every instance of the word "former" with precise start and end times (seconds). Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
6.4,
6.8100000000000005,
"former"
]
] | offline | [{"word":"former","start":6.4,"end":6.81}] |
3853-163249-0054 | dev-clean | LibriSpeech/dev-clean/3853/163249/3853-163249-0054.flac | semantic_detection | body_part | Do not answer until the audio ends. Then list each instance of a body part as {"word", "start", "end"}. If none, reply []. | [
[
1.82,
2.26,
"breath"
],
[
11.36,
11.81,
"hearts"
],
[
16.53,
16.84,
"blood"
]
] | offline | [{"word":"breath","start":1.82,"end":2.26},{"word":"hearts","start":11.36,"end":11.81},{"word":"blood","start":16.53,"end":16.84}] |
2078-142845-0026 | dev-clean | LibriSpeech/dev-clean/2078/142845/2078-142845-0026.flac | lexical_detection | toast | After the audio ends, extract all occurrences of the word "toast". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.44,
2.2,
"toast"
]
] | offline | [{"word":"toast","start":1.44,"end":2.2}] |
2803-154320-0012 | dev-clean | LibriSpeech/dev-clean/2803/154320/2803-154320-0012.flac | lexical_detection | brute | Once the audio has finished, identify every instance of the word "brute". Return a JSON list of {"word", "start", "end"} in time order. If none, reply []. | [
[
1.82,
2.55,
"brute"
]
] | offline | [{"word":"brute","start":1.82,"end":2.55}] |
8842-304647-0000 | dev-clean | LibriSpeech/dev-clean/8842/304647/8842-304647-0000.flac | lexical_detection | limb | Listen to the full audio. After it finishes, find every instance of the word "limb". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
3.14,
3.49,
"limb"
]
] | offline | [{"word":"limb","start":3.14,"end":3.49}] |
3081-166546-0019 | dev-clean | LibriSpeech/dev-clean/3081/166546/3081-166546-0019.flac | semantic_detection | person_name | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.48,
1.83,
"george"
]
] | offline | [{"word":"george","start":1.48,"end":1.83}] |
3752-4943-0025 | dev-clean | LibriSpeech/dev-clean/3752/4943/3752-4943-0025.flac | semantic_detection | person_name | 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 []. | [
[
0.16,
0.63,
"kirkland"
]
] | offline | [{"word":"kirkland","start":0.16,"end":0.63}] |
3081-166546-0067 | dev-clean | LibriSpeech/dev-clean/3081/166546/3081-166546-0067.flac | lexical_detection | glimpse | You will hear a complete utterance. When it ends, list every instance of the word "glimpse" you heard with word-level timestamps. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
6.5,
6.98,
"glimpse"
]
] | offline | [{"word":"glimpse","start":6.5,"end":6.98}] |
1272-141231-0030 | dev-clean | LibriSpeech/dev-clean/1272/141231/1272-141231-0030.flac | semantic_detection | body_part | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
3.26,
3.68,
"face"
]
] | offline | [{"word":"face","start":3.26,"end":3.68}] |
3081-166546-0053 | dev-clean | LibriSpeech/dev-clean/3081/166546/3081-166546-0053.flac | semantic_detection | ordinal | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.5899999999999999,
2.18,
"first"
]
] | offline | [{"word":"first","start":1.59,"end":2.18}] |
1272-135031-0000 | dev-clean | LibriSpeech/dev-clean/1272/135031/1272-135031-0000.flac | semantic_detection | animal | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
9.89,
10.38,
"dove"
]
] | offline | [{"word":"dove","start":9.89,"end":10.38}] |
1993-147149-0006 | dev-clean | LibriSpeech/dev-clean/1993/147149/1993-147149-0006.flac | semantic_counting | person_name | 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 []. | [
[
3.31,
3.9699999999999998,
"one"
],
[
17.5,
17.99,
"two"
]
] | offline | [{"word":"one","start":3.31,"end":3.97},{"word":"two","start":17.5,"end":17.99}] |
3853-163249-0026 | dev-clean | LibriSpeech/dev-clean/3853/163249/3853-163249-0026.flac | lexical_detection | gifts | You will hear a complete utterance. When it ends, list every instance of the word "gifts" you heard with word-level timestamps. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
3.05,
3.39,
"gifts"
]
] | offline | [{"word":"gifts","start":3.05,"end":3.39}] |
1462-170142-0012 | dev-clean | LibriSpeech/dev-clean/1462/170142/1462-170142-0012.flac | lexical_detection | shadows | After listening to the whole recording, output every instance of the word "shadows" with precise start and end times (seconds). Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
4.43,
4.92,
"shadows"
]
] | offline | [{"word":"shadows","start":4.43,"end":4.92}] |
6345-93306-0023 | dev-clean | LibriSpeech/dev-clean/6345/93306/6345-93306-0023.flac | lexical_detection | certainly | Once the audio has finished, identify every instance of the word "certainly". Return a JSON list of {"word", "start", "end"} in time order. If none, reply []. | [
[
1.46,
1.9100000000000001,
"certainly"
]
] | offline | [{"word":"certainly","start":1.46,"end":1.91}] |
3170-137482-0007 | dev-clean | LibriSpeech/dev-clean/3170/137482/3170-137482-0007.flac | lexical_detection | fun | Process the full audio offline. Find every instance of the word "fun" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
4.44,
4.72,
"fun"
]
] | offline | [{"word":"fun","start":4.44,"end":4.72}] |
7976-110523-0015 | dev-clean | LibriSpeech/dev-clean/7976/110523/7976-110523-0015.flac | lexical_detection | grethel | After listening to the whole recording, output every instance of the word "grethel" with precise start and end times (seconds). Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
0.4,
0.75,
"grethel"
]
] | offline | [{"word":"grethel","start":0.4,"end":0.75}] |
6295-64301-0017 | dev-clean | LibriSpeech/dev-clean/6295/64301/6295-64301-0017.flac | semantic_detection | person_name | 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 []. | [
[
0.21,
0.54,
"letty"
]
] | offline | [{"word":"letty","start":0.21,"end":0.54}] |
3576-138058-0000 | dev-clean | LibriSpeech/dev-clean/3576/138058/3576-138058-0000.flac | semantic_detection | person_name | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
6.68,
7.37,
"sancho"
]
] | offline | [{"word":"sancho","start":6.68,"end":7.37}] |
5694-64025-0005 | dev-clean | LibriSpeech/dev-clean/5694/64025/5694-64025-0005.flac | semantic_detection | organization | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
5.24,
5.52,
"army"
]
] | offline | [{"word":"army","start":5.24,"end":5.52}] |
3081-166546-0052 | dev-clean | LibriSpeech/dev-clean/3081/166546/3081-166546-0052.flac | lexical_detection | musicians | When the utterance is complete, report all instances of the word "musicians" including the spoken word and time span. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.4,
2.07,
"musicians"
]
] | offline | [{"word":"musicians","start":1.4,"end":2.07}] |
6313-66129-0015 | dev-clean | LibriSpeech/dev-clean/6313/66129/6313-66129-0015.flac | semantic_detection | animal | 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 []. | [
[
3.52,
3.99,
"ponies"
]
] | offline | [{"word":"ponies","start":3.52,"end":3.99}] |
2428-83699-0005 | dev-clean | LibriSpeech/dev-clean/2428/83699/2428-83699-0005.flac | lexical_counting | nothing | Process the full audio offline. For every occurrence of the word "nothing", emit a count-word event with timestamps. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
0.45,
0.76,
"one"
],
[
3.94,
4.2,
"two"
]
] | offline | [{"word":"one","start":0.45,"end":0.76},{"word":"two","start":3.94,"end":4.2}] |
6313-66129-0010 | dev-clean | LibriSpeech/dev-clean/6313/66129/6313-66129-0010.flac | lexical_detection | reckon | Do not answer until the audio ends. Then list each instance of the word "reckon" as {"word", "start", "end"}. If none, reply []. | [
[
0.6000000000000001,
0.93,
"reckon"
]
] | offline | [{"word":"reckon","start":0.6,"end":0.93}] |
3752-4944-0047 | dev-clean | LibriSpeech/dev-clean/3752/4944/3752-4944-0047.flac | semantic_detection | body_part | 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 []. | [
[
2.9699999999999998,
3.33,
"face"
],
[
6.33,
6.72,
"head"
]
] | offline | [{"word":"face","start":2.97,"end":3.33},{"word":"head","start":6.33,"end":6.72}] |
1919-142785-0003 | dev-clean | LibriSpeech/dev-clean/1919/142785/1919-142785-0003.flac | semantic_counting | plant | Once the audio has finished, list each occurrence of a plant name as {"word": <count>, "start", "end"} in order. If none, reply []. | [
[
0.5,
1.17,
"one"
],
[
2.46,
2.86,
"two"
]
] | offline | [{"word":"one","start":0.5,"end":1.17},{"word":"two","start":2.46,"end":2.86}] |
8842-304647-0003 | dev-clean | LibriSpeech/dev-clean/8842/304647/8842-304647-0003.flac | semantic_detection | body_part | 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 []. | [
[
2.32,
2.68,
"eyes"
]
] | offline | [{"word":"eyes","start":2.32,"end":2.68}] |
2078-142845-0049 | dev-clean | LibriSpeech/dev-clean/2078/142845/2078-142845-0049.flac | lexical_detection | rusks | Listen carefully to the entire clip, then report every instance of the word "rusks" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.62,
2.37,
"rusks"
]
] | offline | [{"word":"rusks","start":1.62,"end":2.37}] |
2803-154328-0014 | dev-clean | LibriSpeech/dev-clean/2803/154328/2803-154328-0014.flac | semantic_detection | profession | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
2.18,
2.66,
"major"
]
] | offline | [{"word":"major","start":2.18,"end":2.66}] |
3752-4943-0009 | dev-clean | LibriSpeech/dev-clean/3752/4943/3752-4943-0009.flac | lexical_detection | order | After listening to the whole recording, output every instance of the word "order" with precise start and end times (seconds). Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
0.45,
0.74,
"order"
]
] | offline | [{"word":"order","start":0.45,"end":0.74}] |
2086-149220-0037 | dev-clean | LibriSpeech/dev-clean/2086/149220/2086-149220-0037.flac | semantic_detection | kinship | 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 []. | [
[
1.27,
1.58,
"cousin"
]
] | offline | [{"word":"cousin","start":1.27,"end":1.58}] |
2086-149220-0013 | dev-clean | LibriSpeech/dev-clean/2086/149220/2086-149220-0013.flac | lexical_detection | antique | Process the full audio offline. Find every instance of the word "antique" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.72,
2.47,
"antique"
]
] | offline | [{"word":"antique","start":1.72,"end":2.47}] |
6313-66125-0001 | dev-clean | LibriSpeech/dev-clean/6313/66125/6313-66125-0001.flac | semantic_detection | plant | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.27,
1.7000000000000002,
"bushes"
]
] | offline | [{"word":"bushes","start":1.27,"end":1.7}] |
5536-43358-0001 | dev-clean | LibriSpeech/dev-clean/5536/43358/5536-43358-0001.flac | lexical_detection | eternal | Listen carefully to the entire clip, then report every instance of the word "eternal" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
2.92,
3.5700000000000003,
"eternal"
]
] | offline | [{"word":"eternal","start":2.92,"end":3.57}] |
422-122949-0018 | dev-clean | LibriSpeech/dev-clean/422/122949/422-122949-0018.flac | semantic_detection | body_part | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
13,
13.37,
"body"
]
] | offline | [{"word":"body","start":13.0,"end":13.37}] |
6319-64726-0000 | dev-clean | LibriSpeech/dev-clean/6319/64726/6319-64726-0000.flac | semantic_detection | food | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
6.45,
6.93,
"feast"
]
] | offline | [{"word":"feast","start":6.45,"end":6.93}] |
3853-163249-0046 | dev-clean | LibriSpeech/dev-clean/3853/163249/3853-163249-0046.flac | lexical_detection | fails | Do not answer until the audio ends. Then list each instance of the word "fails" as {"word", "start", "end"}. If none, reply []. | [
[
10.21,
10.71,
"fails"
]
] | offline | [{"word":"fails","start":10.21,"end":10.71}] |
7850-281318-0008 | dev-clean | LibriSpeech/dev-clean/7850/281318/7850-281318-0008.flac | lexical_detection | birds | Listen carefully to the entire clip, then report every instance of the word "birds" as JSON objects with word, start, and end. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.6099999999999999,
2.06,
"birds"
]
] | offline | [{"word":"birds","start":1.61,"end":2.06}] |
652-129742-0005 | dev-clean | LibriSpeech/dev-clean/652/129742/652-129742-0005.flac | semantic_detection | plant | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
2.56,
3.04,
"cabbage"
]
] | offline | [{"word":"cabbage","start":2.56,"end":3.04}] |
3576-138058-0001 | dev-clean | LibriSpeech/dev-clean/3576/138058/3576-138058-0001.flac | semantic_detection | body_part | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
5.94,
6.25,
"eye"
]
] | offline | [{"word":"eye","start":5.94,"end":6.25}] |
3081-166546-0082 | dev-clean | LibriSpeech/dev-clean/3081/166546/3081-166546-0082.flac | semantic_counting | cardinal | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
2.06,
2.52,
"one"
],
[
3.68,
4.29,
"two"
]
] | offline | [{"word":"one","start":2.06,"end":2.52},{"word":"two","start":3.68,"end":4.29}] |
5895-34629-0009 | dev-clean | LibriSpeech/dev-clean/5895/34629/5895-34629-0009.flac | semantic_detection | city | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
0.89,
1.35,
"london"
]
] | offline | [{"word":"london","start":0.89,"end":1.35}] |
7850-286674-0013 | dev-clean | LibriSpeech/dev-clean/7850/286674/7850-286674-0013.flac | lexical_detection | comes | After the audio ends, extract all occurrences of the word "comes". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
0.36,
0.64,
"comes"
]
] | offline | [{"word":"comes","start":0.36,"end":0.64}] |
8297-275154-0014 | dev-clean | LibriSpeech/dev-clean/8297/275154/8297-275154-0014.flac | lexical_detection | dressed | Listen to the full audio. After it finishes, find every instance of the word "dressed". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
6.25,
6.88,
"dressed"
]
] | offline | [{"word":"dressed","start":6.25,"end":6.88}] |
1673-143397-0010 | dev-clean | LibriSpeech/dev-clean/1673/143397/1673-143397-0010.flac | semantic_counting | profession | 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": <string>, "start": <seconds>, "end": <seconds>}... | [
[
4.57,
4.91,
"one"
],
[
5.92,
6.32,
"two"
]
] | offline | [{"word":"one","start":4.57,"end":4.91},{"word":"two","start":5.92,"end":6.32}] |
2086-149220-0010 | dev-clean | LibriSpeech/dev-clean/2086/149220/2086-149220-0010.flac | semantic_detection | food | 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 []. | [
[
4.05,
4.52,
"egg"
]
] | offline | [{"word":"egg","start":4.05,"end":4.52}] |
6319-57405-0010 | dev-clean | LibriSpeech/dev-clean/6319/57405/6319-57405-0010.flac | semantic_detection | person_name | 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 []. | [
[
2.48,
2.9,
"pyrrha"
]
] | offline | [{"word":"pyrrha","start":2.48,"end":2.9}] |
1673-143396-0001 | dev-clean | LibriSpeech/dev-clean/1673/143396/1673-143396-0001.flac | lexical_detection | disappeared | After the audio ends, extract all occurrences of the word "disappeared". Output each spoken word with its start and end time. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.28,
2,
"disappeared"
]
] | offline | [{"word":"disappeared","start":1.28,"end":2.0}] |
6241-66616-0008 | dev-clean | LibriSpeech/dev-clean/6241/66616/6241-66616-0008.flac | semantic_detection | kinship | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
3.88,
4.6,
"forefathers"
],
[
8.89,
9.3,
"mother"
],
[
16.58,
17.03,
"father's"
]
] | offline | [{"word":"forefathers","start":3.88,"end":4.6},{"word":"mother","start":8.89,"end":9.3},{"word":"father's","start":16.58,"end":17.03}] |
8297-275155-0020 | dev-clean | LibriSpeech/dev-clean/8297/275155/8297-275155-0020.flac | semantic_detection | kinship | Do not answer until the audio ends. Then list each instance of a family relation as {"word", "start", "end"}. If none, reply []. | [
[
6.15,
6.86,
"mother"
]
] | offline | [{"word":"mother","start":6.15,"end":6.86}] |
1272-135031-0005 | dev-clean | LibriSpeech/dev-clean/1272/135031/1272-135031-0005.flac | semantic_detection | person_name | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
0.9,
1.3900000000000001,
"ruggedo"
]
] | offline | [{"word":"ruggedo","start":0.9,"end":1.39}] |
777-126732-0056 | dev-clean | LibriSpeech/dev-clean/777/126732/777-126732-0056.flac | semantic_detection | color | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
3.14,
3.43,
"white"
],
[
6.79,
7.04,
"dark"
]
] | offline | [{"word":"white","start":3.14,"end":3.43},{"word":"dark","start":6.79,"end":7.04}] |
2035-152373-0010 | dev-clean | LibriSpeech/dev-clean/2035/152373/2035-152373-0010.flac | semantic_detection | country | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.13,
1.81,
"northumbria"
]
] | offline | [{"word":"northumbria","start":1.13,"end":1.81}] |
5694-64029-0031 | dev-clean | LibriSpeech/dev-clean/5694/64029/5694-64029-0031.flac | lexical_detection | fired | Do not answer until the audio ends. Then list each instance of the word "fired" as {"word", "start", "end"}. If none, reply []. | [
[
5.21,
5.61,
"fired"
]
] | offline | [{"word":"fired","start":5.21,"end":5.61}] |
8842-304647-0005 | dev-clean | LibriSpeech/dev-clean/8842/304647/8842-304647-0005.flac | semantic_detection | animal | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.22,
1.78,
"bestia"
]
] | offline | [{"word":"bestia","start":1.22,"end":1.78}] |
1272-141231-0010 | dev-clean | LibriSpeech/dev-clean/1272/141231/1272-141231-0010.flac | semantic_detection | cardinal | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
3.83,
4.08,
"two"
],
[
6.79,
7.47,
"thousands"
]
] | offline | [{"word":"two","start":3.83,"end":4.08},{"word":"thousands","start":6.79,"end":7.47}] |
2902-9006-0002 | dev-clean | LibriSpeech/dev-clean/2902/9006/2902-9006-0002.flac | lexical_detection | centuries | Once the audio has finished, identify every instance of the word "centuries". Return a JSON list of {"word", "start", "end"} in time order. If none, reply []. | [
[
10.76,
11.49,
"centuries"
]
] | offline | [{"word":"centuries","start":10.76,"end":11.49}] |
5536-43358-0017 | dev-clean | LibriSpeech/dev-clean/5536/43358/5536-43358-0017.flac | lexical_counting | mystery | After listening to the whole recording, output timed count words for the word "mystery" (one, two, three, ...) as JSON. If none, reply []. | [
[
1.2,
1.62,
"one"
],
[
6.21,
6.64,
"two"
]
] | offline | [{"word":"one","start":1.2,"end":1.62},{"word":"two","start":6.21,"end":6.64}] |
2277-149874-0011 | dev-clean | LibriSpeech/dev-clean/2277/149874/2277-149874-0011.flac | lexical_detection | thinking | Listen to the audio in full, then emit a JSON array of every instance of the word "thinking" with fields word, start, end (seconds). If none, reply []. | [
[
0.88,
1.22,
"thinking"
]
] | offline | [{"word":"thinking","start":0.88,"end":1.22}] |
652-130737-0005 | dev-clean | LibriSpeech/dev-clean/652/130737/652-130737-0005.flac | semantic_detection | food | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
0.49,
1.02,
"clarets"
]
] | offline | [{"word":"clarets","start":0.49,"end":1.02}] |
1919-142785-0059 | dev-clean | LibriSpeech/dev-clean/1919/142785/1919-142785-0059.flac | semantic_detection | cardinal | Do not answer until the audio ends. Then list each instance of a number as {"word", "start", "end"}. If none, reply []. | [
[
1.9,
2.16,
"one"
]
] | offline | [{"word":"one","start":1.9,"end":2.16}] |
251-118436-0017 | dev-clean | LibriSpeech/dev-clean/251/118436/251-118436-0017.flac | lexical_detection | fingers | Listen to the full audio. After it finishes, find every instance of the word "fingers". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
0.48,
1.07,
"fingers"
]
] | offline | [{"word":"fingers","start":0.48,"end":1.07}] |
2803-154328-0003 | dev-clean | LibriSpeech/dev-clean/2803/154328/2803-154328-0003.flac | semantic_detection | body_part | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
5.7,
6.45,
"eyelids"
],
[
8.28,
8.8,
"arms"
]
] | offline | [{"word":"eyelids","start":5.7,"end":6.45},{"word":"arms","start":8.28,"end":8.8}] |
2803-154328-0021 | dev-clean | LibriSpeech/dev-clean/2803/154328/2803-154328-0021.flac | semantic_detection | person_name | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
2.43,
2.81,
"robert"
]
] | offline | [{"word":"robert","start":2.43,"end":2.81}] |
1919-142785-0062 | dev-clean | LibriSpeech/dev-clean/1919/142785/1919-142785-0062.flac | semantic_detection | food | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
3.23,
3.63,
"salmon"
]
] | offline | [{"word":"salmon","start":3.23,"end":3.63}] |
777-126732-0027 | dev-clean | LibriSpeech/dev-clean/777/126732/777-126732-0027.flac | lexical_detection | skin | You will hear a complete utterance. When it ends, list every instance of the word "skin" you heard with word-level timestamps. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
0.66,
1,
"skin"
]
] | offline | [{"word":"skin","start":0.66,"end":1.0}] |
174-168635-0019 | dev-clean | LibriSpeech/dev-clean/174/168635/174-168635-0019.flac | semantic_detection | person_name | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.62,
2.58,
"jean valjean"
]
] | offline | [{"word":"jean valjean","start":1.62,"end":2.58}] |
251-136532-0018 | dev-clean | LibriSpeech/dev-clean/251/136532/251-136532-0018.flac | lexical_detection | question | Once the audio has finished, identify every instance of the word "question". Return a JSON list of {"word", "start", "end"} in time order. If none, reply []. | [
[
0.72,
1.38,
"question"
]
] | offline | [{"word":"question","start":0.72,"end":1.38}] |
2902-9008-0005 | dev-clean | LibriSpeech/dev-clean/2902/9008/2902-9008-0005.flac | lexical_detection | cast | Process the full audio offline. Find every instance of the word "cast" and give the surface word plus start/end timestamps. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
0.86,
1.29,
"cast"
],
[
3.51,
3.86,
"cast"
]
] | offline | [{"word":"cast","start":0.86,"end":1.29},{"word":"cast","start":3.51,"end":3.86}] |
2086-149220-0005 | dev-clean | LibriSpeech/dev-clean/2086/149220/2086-149220-0005.flac | lexical_detection | bees | When the utterance is complete, report all instances of the word "bees" including the spoken word and time span. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
0.5,
0.84,
"bees"
]
] | offline | [{"word":"bees","start":0.5,"end":0.84}] |
1993-147149-0009 | dev-clean | LibriSpeech/dev-clean/1993/147149/1993-147149-0009.flac | lexical_detection | upstairs | Listen to the audio in full, then emit a JSON array of every instance of the word "upstairs" with fields word, start, end (seconds). If none, reply []. | [
[
14.03,
14.66,
"upstairs"
]
] | offline | [{"word":"upstairs","start":14.03,"end":14.66}] |
2803-161169-0005 | dev-clean | LibriSpeech/dev-clean/2803/161169/2803-161169-0005.flac | semantic_detection | cardinal | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
10.38,
10.61,
"one"
]
] | offline | [{"word":"one","start":10.38,"end":10.61}] |
3536-8226-0005 | dev-clean | LibriSpeech/dev-clean/3536/8226/3536-8226-0005.flac | semantic_detection | profession | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.9,
2.21,
"judge"
],
[
11.31,
11.85,
"lawyers"
]
] | offline | [{"word":"judge","start":1.9,"end":2.21},{"word":"lawyers","start":11.31,"end":11.85}] |
6241-66616-0005 | dev-clean | LibriSpeech/dev-clean/6241/66616/6241-66616-0005.flac | lexical_detection | generation | After listening to the whole recording, output every instance of the word "generation" with precise start and end times (seconds). Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
9.76,
10.51,
"generation"
]
] | offline | [{"word":"generation","start":9.76,"end":10.51}] |
3000-15664-0016 | dev-clean | LibriSpeech/dev-clean/3000/15664/3000-15664-0016.flac | semantic_detection | color | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.28,
1.6800000000000002,
"emerald"
]
] | offline | [{"word":"emerald","start":1.28,"end":1.68}] |
2412-153954-0003 | dev-clean | LibriSpeech/dev-clean/2412/153954/2412-153954-0003.flac | semantic_detection | plant | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
0.56,
0.9,
"vine"
]
] | offline | [{"word":"vine","start":0.56,"end":0.9}] |
8297-275156-0003 | dev-clean | LibriSpeech/dev-clean/8297/275156/8297-275156-0003.flac | lexical_counting | gossip | Listen to the audio in full, then emit a JSON array of count events for the word "gossip" (word, start, end). If none, reply []. | [
[
1.75,
2.22,
"one"
],
[
5.38,
5.85,
"two"
]
] | offline | [{"word":"one","start":1.75,"end":2.22},{"word":"two","start":5.38,"end":5.85}] |
1993-147964-0005 | dev-clean | LibriSpeech/dev-clean/1993/147964/1993-147964-0005.flac | semantic_detection | animal | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.38,
1.71,
"horse"
]
] | offline | [{"word":"horse","start":1.38,"end":1.71}] |
2035-147961-0005 | dev-clean | LibriSpeech/dev-clean/2035/147961/2035-147961-0005.flac | lexical_detection | armies | Listen to the full audio. After it finishes, find every instance of the word "armies". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.87,
2.39,
"armies"
]
] | offline | [{"word":"armies","start":1.87,"end":2.39}] |
777-126732-0007 | dev-clean | LibriSpeech/dev-clean/777/126732/777-126732-0007.flac | semantic_detection | body_part | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
3.79,
4.18,
"hand"
]
] | offline | [{"word":"hand","start":3.79,"end":4.18}] |
652-129742-0009 | dev-clean | LibriSpeech/dev-clean/652/129742/652-129742-0009.flac | lexical_counting | cups | After listening to the whole recording, output timed count words for the word "cups" (one, two, three, ...) as JSON. If none, reply []. | [
[
5.02,
5.53,
"one"
],
[
10.07,
10.51,
"two"
]
] | offline | [{"word":"one","start":5.02,"end":5.53},{"word":"two","start":10.07,"end":10.51}] |
3853-163249-0002 | dev-clean | LibriSpeech/dev-clean/3853/163249/3853-163249-0002.flac | semantic_detection | body_part | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.77,
2,
"ear"
]
] | offline | [{"word":"ear","start":1.77,"end":2.0}] |
5338-284437-0020 | dev-clean | LibriSpeech/dev-clean/5338/284437/5338-284437-0020.flac | semantic_detection | profession | 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": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
6.74,
7.2,
"servant"
]
] | offline | [{"word":"servant","start":6.74,"end":7.2}] |
1988-24833-0015 | dev-clean | LibriSpeech/dev-clean/1988/24833/1988-24833-0015.flac | lexical_detection | upset | Listen to the full audio. After it finishes, find every instance of the word "upset". For each match, report the spoken word and its start/end times in seconds. Reply with JSON only: a list of objects {"word": <string>, "start": <seconds>, "end": <seconds>} in chronological order. If none, reply []. | [
[
1.47,
1.85,
"upset"
]
] | offline | [{"word":"upset","start":1.47,"end":1.85}] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.