Verified70% confidenceFactExact time
Research 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.
4
Sources
70%
Confidence
Long-term
Relevance
7/7/2026
First Seen
Sources
AI超级预测者：机器能否超越人类专家的预测能力？
hackernewshackernews7/6/2026
Related Claims
UnverifiedAcademia 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.84% similarUnverified思维链推理的效果随模型规模呈现涌现性，在较小模型上几乎无效，在足够大的模型上才能显著提升推理准确率84% similarUnverified多项实验表明，模型的表面推理与其实际决策机制之间存在显著偏差，通过激活修补可发现真正影响输出的内部路径与书面推理不吻合80% similarUnverifiedHidden Chains of Thought do not require explicitly displaying each step of the thinking process in the output, allowing the model to perform reasoning iterations within its internal state space.79% similarUnverifiedModel Routing and Hybrid Inference have already been explored in academia and were not pioneered by Cognition.79% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/201918API
curl https://kongchang.com/api/v1/knowledge/claims/201918MCP
get_claim(id=201918)