method
active
method:reading-chain-of-thoughtReading chain-of-thought
Behavior-based explanation that is not fully faithful.
Neighborhood — ranked by edge-count
Papers (1)
paper
Methods (1)
method
- Chain-of-thought promptingrelated_toTechnique by which LLMs generate intermediate reasoning steps before final output; used by ChatGPT o3.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- A technique that outputs intermediate reasoning steps, used here to detect verbalized eval awareness.
- Medium through which eval awareness is often verbalized; target of intervention.
- Monitoring reasoning traces for misbehavior.
- Central concept: verbalized reasoning that occurs after the model has already internally settled on an answer, particularly on easier tasks.
- Whether written reasoning reflects actual model mechanisms.
- A prompting technique that elicits intermediate reasoning steps before final answer inference in language models.
- Using multi-step reasoning by generating intermediate thoughts.
- Mechanism in which reasoning models consult safety policies within their chain-of-thought before answering, documented for gpt-oss