ResearchBenchmarksTheory of Mind Question Answering on FanToMFollow91.7AccuracyHuman53.84463.67273.583.328Apr 20, 2026Evaluation ResultsMethodMethodLinksAccuracyHumanEvaluator LLM=GPT-4oEvaluator LLM=GPT-4o2026.0491.7TTEvaluator LLM=GPT-4oEvaluator LLM=GPT-4o2026.0487.4PDDL-MINDEvaluator LLM=GPT-4oEvaluator LLM=GPT-4o2026.0480EnigmaToMEvaluator LLM=GPT-4oEvaluator LLM=GPT-4o2026.0455.3