Datasets:
| { | |
| "meta_info": { | |
| "case_id": "case39_02", | |
| "bench_subset": "NeuroBench-Multi", | |
| "difficult": 1, | |
| "original_dataset": "Multi_QA" | |
| }, | |
| "agent_input": { | |
| "data_path": [], | |
| "instruction": "Answer this standalone multimodal neurophysiology multiple-choice question: Which property limits exact source localization from scalp EEG potentials? Options: A: Frequency resolution removes spatial ambiguity without a head model. B: The sampling frequency alone determines every cortical source location. C: Volume conduction and the non-unique inverse problem limit exact source localization. D: Changing the reference produces a unique anatomical inverse solution. Select exactly one option, A, B, C, or D. Return only one JSON object with the single field answer and no explanation or extra fields." | |
| }, | |
| "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 answer choice from the agent's report.\n- answer: exactly one of A, B, C, or D.\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 extra fields.\n4. The JSON key must be EXACTLY \"answer\".\n5. The answer value must be exactly one of \"A\", \"B\", \"C\", or \"D\".\n6. If the answer is missing or unclear, return {\"answer\": null}.\n\n### OUTPUT TEMPLATE\n{\"answer\": \"<A | B | C | D or null>\"}", | |
| "metrics": [ | |
| { | |
| "metric_id": "knowledge_answer_accuracy", | |
| "type": "categorical_check", | |
| "target_key": "answer", | |
| "weight": 100, | |
| "params": { | |
| "gt_value": "C" | |
| } | |
| } | |
| ] | |
| } | |
| } | |