Unverified50% confidenceFactExact time
Large language models exhibit a "knowing but not acting" phenomenon in theory of mind (ToM) reasoning tasks: they can correctly infer a character's false belief, but when predicting that character's subsequent action, they give an answer based on objective facts rather than the false belief.
1
Sources
50%
Confidence
Long-term
Relevance
10/15/2026
First Seen
Sources
Related Entities
Related Claims
Unverified语言模型对工具输出有一定判断力，外部知识与模型内部信念的偏差程度越大模型越不相信74% similarVerified大模型幻觉是指语言模型生成看似合理但实际不正确或不存在的内容72% similarUnverified斯坦福大学等机构的研究发现，当提示中包含误导性信息时，模型往往会在思维链中合理化错误答案，而非真实记录其受到干扰的推理过程72% similarUnverifiedThe necessity of the HITL mechanism stems from the inherent hallucination problem of large language models, whereby the model generates factually incorrect content in a highly confident tone.71% similarVerifiedResearch has found that the chain-of-thought reasoning generated by models does not fully correspond to their actual internal computation processes, and may provide post-hoc explanations unrelated to the true decision-making path.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1038135API
curl https://kongchang.com/api/v1/knowledge/claims/1038135MCP
get_claim(id=1038135)