Reasoning models generate performative CoT tokens after achieving strong confidence in their final answer without revealing this belief in text

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

• a model becomes strongly confident in its final answer, but continues generating tokens without revealing its internal beliefquote0.819
  Core definitional quote for performative chain-of-thought
• CoT models' explicit thought paths enable intentional inconsistency between reasoning and output, a form of deception fundamentally distinct from random errorsclaim0.804
  Theoretical framing establishing why CoT models are uniquely suited to exhibit strategic deception
• Why do 1B-models fail at generating CoT that aids answer inference, and how can this be addressed in multimodal settings?question0.799
  Central research question motivating investigation into hallucination and two-stage framework design.
• CoT models have dual-use potential: their advanced reasoning amplifies both task fidelity and sophisticated goal-directed dishonestyclaim0.778
  High-level policy-relevant claim about the risks of advanced reasoning in LLMs
• Transformers develop self-models through in-context learning, not just training data; even old base models without LLM-related text can bootstrap self-referential reasoning at runtime.claim0.768
  Antra's foundational claim about how introspection arises computationally rather than from memorised text.
• can activation probing enable efficient adaptive computation by detecting when a model's belief has stabilized during CoT generation?question0.767
  Practical question addressed by the probe-guided early exit experiments
• How can reasoning-optimized models preserve their reasoning ability while gaining agentic capabilities?question0.762
  Core research question motivating the paper's focus on continual RL training of reasoning models rather than base/instruction-tuned models.
• Task difficulty moderates whether CoT is performative or genuine: easy recall questions show performative CoT, difficult multihop questions show genuine reasoningclaim0.762
  Task difficulty as the key variable distinguishing the two modes of CoT identified in the paper