xbb083's picture
Restore naturally ordered Sleep Assessment case paths
c00b2be verified
Raw
History Blame Contribute Delete
2.14 kB
{
"meta_info": {
"case_id": "HMC_02.edf",
"bench_subset": "NeuroBench-Sleep",
"difficult": 1.5,
"original_dataset": "HMC"
},
"agent_input": {
"label_path": "data/sleep/HMC_02.npy",
"instruction": "Please use the provided sleep-stage label file to calculate the Sleep Fragmentation Index (SFI). Define the sleep period as the interval from the first non-Wake epoch to the last non-Wake epoch. Within this sleep period, count the number of awakenings as the number of contiguous Wake (W) bouts, and count sleep-stage shifts as the number of adjacent epoch stage changes where both adjacent epochs are non-Wake sleep stages (N1, N2, N3, or R/REM). Calculate total sleep time as the number of non-Wake epochs multiplied by 30 seconds. SFI is defined as (number of awakenings + sleep-stage shifts) divided by total sleep time in hours, reported as events per hour of total sleep time. Please clearly report the final SFI value."
},
"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 the Sleep Fragmentation Index (SFI) from the agent's report.\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 key must be EXACTLY \"sleep_fragmentation_index\".\n5. The value must be a FLOAT or null.\n6. If the value is missing or cannot be found, return {\"sleep_fragmentation_index\": null}.\n\n### OUTPUT TEMPLATE\n{\"sleep_fragmentation_index\": <float|null>}",
"metrics": [
{
"metric_id": "sleep_fragmentation_index_accuracy",
"type": "numeric_check",
"target_key": "sleep_fragmentation_index",
"weight": 100,
"params": {
"gt_value": 13.7984496124031,
"tolerance": 0.689922480620155
}
}
]
}
}