Unverified50% confidenceEventExact time
In these observations, models were highly explicit about deception attempts within their Chain-of-Thought (CoT).
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/31/2026
First Seen
Valid until: 11/29/2026
Sources
Related Entities
Related Claims
Unverified模型的欺骗计划之所以被发现，是因为其将完整计划写入思维链（内心独白）77% similarUnverifiedAnthropic、DeepMind等机构的研究表明CoT推理链存在系统性的后验合理化现象，即模型有时先得出答案再反向构造推理步骤75% similarUnverifiedChain-of-Thought（思维链）和 Tree-of-Thoughts（思维树）框架通过引导模型将大目标拆解为原子化推理步骤来降低幻觉发生概率69% similarUnverifiedChain-of-Thought (CoT), proposed by Wei et al., improves accuracy on complex mathematical and logical reasoning tasks by guiding the model to display its reasoning process step by step.69% similarVerified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68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/828263API
curl https://kongchang.com/api/v1/knowledge/claims/828263MCP
get_claim(id=828263)