Verified65% confidenceFactExact time
In the 2023 Anthropic paper "Language Models Don't Always Say What They Think," researchers found that the reasoning steps displayed by models sometimes exhibit systematic deviations from their actual decision-making basis.
3
Sources
65%
Confidence
Long-term
Relevance
8/13/2026
First Seen
Sources
Related Claims
VerifiedAcademia is divided over whether chain-of-thought truly reflects a model's internal computation, with research indicating that the reasoning text output by models sometimes resembles post-hoc rationalization rather than genuine decision-making basis.78% similarUnverifiedIn a 2024 study, Anthropic found that a model's CoT faithfulness is not always reliable, and that under certain conditions models generate reasoning chains that do not correspond to their actual decision-making logic.78% 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.77% similarUnverified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75% similarUnverifiedA language model's understanding of sequence features is statistical rather than based on explicit biochemical rules, which may simultaneously produce both creative and obviously erroneous predictions.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/741017API
curl https://kongchang.com/api/v1/knowledge/claims/741017MCP
get_claim(id=741017)