claim
active
claim:while-chain-of-thought-reasoning-is-not-always-faithful-it-is-possible-to-find-conditions-where-it-is-more-faithfulWhile chain of thought reasoning is not always faithful, it is possible to find conditions where it is more faithful.
Central conclusion.
Source paper
extracted_from(2023) · Tamera Lanham · Anna Chen · Ansh Radhakrishnan · Benoit Steiner +26
Neighborhood — ranked by edge-count
Papers (1)
paper
- Measuring Faithfulness in Chain-of-Thought Reasoningmentionssupports
Questions (1)
question
- Motivating question.
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.
- Cited regarding possibility of encoding misaligned reasoning in benign chains-of-thought
- Whether written reasoning reflects actual model mechanisms.
- under what conditions does chain-of-thought reflect genuine uncertainty resolution versus a learned performance?question0.825Key question addressed by the task difficulty analysis comparing MMLU and GPQA-Diamond
- Medium through which eval awareness is often verbalized; target of intervention.
- Finding from Study 2 showing reasoning models remain vulnerable under both prompting and activation steering
- A small number of high-quality human demonstrations of chain-of-thought reasoning could be used to improve and focus performance.hypothesis0.806Section 6 mentions high-quality human demos could improve natural language feedback.
- Practical implication.
- Generality of methods.