claim
active
claim:reasoning-models-generate-performative-cot-tokens-after-achieving-strong-confidence-in-their-final-answer-without-revealing-this-belief-in-textReasoning models generate performative CoT tokens after achieving strong confidence in their final answer without revealing this belief in text
The central empirical claim of the paper, supported by activation probing evidence
Source paper
extracted_from(2026) · Siddharth Boppana · Annabel Ma · Max Loeffler · Raphaël Sarfati +4
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (1)
finding
- Core empirical result demonstrating early belief formation in easy tasks
Questions (1)
question
- Central research question motivating the paper
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.
- Core definitional quote for performative chain-of-thought
- Tested and rejected.
- For reasoning faithfulness, larger models may behave worse than smaller ones because they can predict answers without CoT.hypothesis0.810Motivates §3.
- Limitations.
- Theoretical framing establishing why CoT models are uniquely suited to exhibit strategic deception
- In motivated reasoning the CoT derives from the human hint, but not in faithful reasoning.hypothesis0.800Tested by varying the suggested answer.
- Central research question motivating investigation into hallucination and two-stage framework design.
- Interpretation of §3.