Unverified70% confidenceFactExact time
Research from institutions such as Anthropic and DeepMind indicates that CoT reasoning chains exhibit a systematic post-hoc rationalization phenomenon, in which models sometimes arrive at an answer first and then construct reasoning steps in reverse.
2
Sources
70%
Confidence
Long-term
Relevance
7/14/2026
First Seen
Sources
潜在推理：超越思维链的下一代AI推理范式
Related Claims
Verified推理能力基于思维链（Chain-of-Thought, CoT）技术，模型在给出最终答案前生成内部逐步推导过程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.78% similarVerifiedReasoning models perform internal Chain-of-Thought before returning final results, consuming large amounts of tokens and computation time in exchange for deeper understanding and lower error rates.76% similarVerifiedBefore ReAct was proposed, the industry mainly relied on Chain-of-Thought (CoT) prompts to guide models in step-by-step reasoning, but CoT's reasoning process could not invoke external tools or perceive execution results midway.76% similarUnverifiedChain-of-Thought (CoT) prompting improves the quality of complex problem solving by requiring the model to reason step by step.76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/507146API
curl https://kongchang.com/api/v1/knowledge/claims/507146MCP
get_claim(id=507146)