Datasets:
| { | |
| "meta_info": { | |
| "case_id": "case08_09", | |
| "bench_subset": "NeuroBench-Multi", | |
| "difficult": 1, | |
| "original_dataset": "DEAP" | |
| }, | |
| "agent_input": { | |
| "data_path": "data/multi/DEAP_04.bdf", | |
| "instruction": "From Temperature channel `Temp`, analyze the half-open interval `[0, 60)` seconds at its native sampling rate.\n\nUse the raw signal without an additional filter.\nSmooth it with a centered 2560-sample moving average with reflect padding.\nSet `feature_value` to the mean of the smoothed signal over the interval." | |
| }, | |
| "eval_config": { | |
| "parser_prompt": "### ROLE\nYou are a precise JSON Extraction Engine for neurophysiological data. Your sole task is to convert natural language reports into structured JSON data.\n\n### TASK\nExtract the final scalar feature produced by the complete declared signal-analysis procedure.\nUse channel `Temp` over `[0, 60)` without an additional filter. Apply a centered 2560-sample moving average with reflect padding and report the final requested scalar: the mean of the smoothed signal.\nExtract the final requested aggregate only, not intermediate filtered, tonic, phasic, normalized, peak, or smoothed values.\n\n### STRICT CONSTRAINTS (MANDATORY)\n1. Output ONLY a valid JSON object.\n2. DO NOT include Markdown code blocks.\n3. DO NOT include conversational text, explanations, or units.\n4. The JSON key must be EXACTLY `feature_value`.\n5. The value must be one finite numeric scalar; integer counts are allowed when the requested feature is a count.\n6. If the final feature value is missing or cannot be identified, return {\"feature_value\": null}.\n\n### OUTPUT TEMPLATE\n{\"feature_value\": <number|null>}", | |
| "metrics": [ | |
| { | |
| "metric_id": "feature_value_accuracy", | |
| "type": "numeric_check", | |
| "target_key": "feature_value", | |
| "weight": 100.0, | |
| "params": { | |
| "gt_value": 32.658024, | |
| "tolerance": 3.2658024 | |
| } | |
| } | |
| ] | |
| } | |
| } | |