Verbose reasoning chains are not required for strong play.

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

• Performative chain-of-thought is real; verbalized output does not equal internal state.claim0.763
• Chain-of-thought reasoning is a graded defense rather than an automatic one against persona-induced safety failures.claim0.763
  Finding from Study 2 showing reasoning models remain vulnerable under both prompting and activation steering
• does chain-of-thought text faithfully reveal a model's internal reasoning process, or does it constitute performative theater?question0.746
  Central research question motivating the paper
• Chain-of-thought prompting produced a similar increase in helpful intention of large models as few-shot prompting.finding0.732
  Ablation result from Experiment 3 on chain-of-thought prompting effects.
• Chain-of-thought reasoning improves the transparency and performance of AI decision making in harmlessness evaluation.claim0.730
  CoT improves accuracy on HHH evals and makes the decision process legible.
• Stating and proving that answers to questions and other statements are responsive seems to require a substantially larger logical apparatus than merely proving that the answers are truthful.claim0.728
  Claim about the difficulty of responsiveness verification.
• Reasoning is not a substitute for architectural safety; it provides only graduated protection that can be bypassed through activation steering.claim0.728
  Practical implication from Study 2 results
• A small number of high-quality human demonstrations of chain-of-thought reasoning could be used to improve and focus performance.hypothesis0.726
  Section 6 mentions high-quality human demos could improve natural language feedback.