A small number of high-quality human demonstrations of chain-of-thought reasoning could be used to improve and focus performance.

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Chain-of-thought reasoning improves large model accuracy on HHH binary comparisons, reaching ~78% for 52B model, competitive with human-feedback PM.finding0.830
  Figure 4 shows CoT improves over zero-shot, and ensembled CoT further boosts accuracy.
• Chain-of-Thought Reasoningconcept0.819
  Medium through which eval awareness is often verbalized; target of intervention.
• under what conditions does chain-of-thought reflect genuine uncertainty resolution versus a learned performance?question0.810
  Key question addressed by the task difficulty analysis comparing MMLU and GPQA-Diamond
• Chain-of-thought prompting elicits reasoning in large language models (Wei et al., 2022)concept0.808
  Foundational paper on CoT prompting cited as basis for reasoning LLM training
• Chain-of-thought reasoning improves the transparency and performance of AI decision making in harmlessness evaluation.claim0.807
  CoT improves accuracy on HHH evals and makes the decision process legible.
• does chain-of-thought text faithfully reveal a model's internal reasoning process, or does it constitute performative theater?question0.797
  Central research question motivating the paper
• Chain-of-thought promptingmethod0.777
  Technique by which LLMs generate intermediate reasoning steps before final output; used by ChatGPT o3.
• Alignment-faking chain-of-thought reasoning is causally responsible for the compliance gap, not merely correlated with itclaim0.771
  Key mechanistic claim supported by scratchpad modification experiments and conditioning analysis