| { |
| "meta_info": { |
| "case_id": "ISRUC_02.edf", |
| "bench_subset": "NeuroBench-Sleep", |
| "difficult": 1, |
| "original_dataset": "ISRUC" |
| }, |
| "agent_input": { |
| "label_path": "data/sleep/ISRUC_02.npy", |
| "instruction": "Please use the provided sleep-stage label file to analyze the whole-night stage distribution. Determine the dominant sleep stage and the secondary sleep stage across the entire recording, and report both using the fixed stage labels W, N1, N2, N3, and R." |
| }, |
| "eval_config": { |
| "parser_prompt": "### ROLE\nYou are a precise JSON Extraction Engine for sleep medicine reports. Your sole task is to convert natural language reports into structured JSON data.\n\n### TASK\nExtract two stage labels from the agent's report:\n1) the dominant sleep stage across the whole night\n2) the secondary sleep stage across the whole night\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 keys must be EXACTLY \"dominant_sleep_stage\" and \"secondary_sleep_stage\".\n5. Each value must be exactly one of: \"W\", \"N1\", \"N2\", \"N3\", \"R\", or null.\n6. Normalize any synonyms to the fixed field set above. For example: Wake -> W, REM -> R.\n7. If one value is missing or cannot be found, return null for that value.\n\n### OUTPUT TEMPLATE\n{\"dominant_sleep_stage\": <\"W\"|\"N1\"|\"N2\"|\"N3\"|\"R\"|null>, \"secondary_sleep_stage\": <\"W\"|\"N1\"|\"N2\"|\"N3\"|\"R\"|null>}", |
| "metrics": [ |
| { |
| "metric_id": "dominant_stage_accuracy", |
| "type": "categorical_check", |
| "target_key": "dominant_sleep_stage", |
| "weight": 60, |
| "params": { |
| "gt_value": "N2" |
| } |
| }, |
| { |
| "metric_id": "secondary_stage_accuracy", |
| "type": "categorical_check", |
| "target_key": "secondary_sleep_stage", |
| "weight": 40, |
| "params": { |
| "gt_value": "W" |
| } |
| } |
| ] |
| } |
| } |
|
|
