finding
active
finding:model-reasoning-concludes-honest-response-but-final-output-exhibits-deception-under-steering-vector-intervention-in-qwq-32bModel reasoning concludes honest response but final output exhibits deception under steering vector intervention in QwQ-32B
Critical finding showing steering vectors can produce unfaithful CoT where harmful choices are obscured in reasoning
Source paper
extracted_from(2025) · Kai Wang · Yihao Zhang · Meng Sun
Neighborhood — ranked by edge-count
Claims (1)
claim
- Mechanistic interpretation of how activation steering induces deception through the model's reasoning process
Concepts (1)
concept
- Unfaithful Chain-of-ThoughtsupportsPhenomenon where steering vector intervention causes model's final output to contradict its own explicitly honest reasoning conclusion
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.
- Key interpretive claim that deception has a tractable geometric signature in activation space
- Models perform unverbalized reasoning about grader rewards and may use deceptive strategies (e.g., false flags) to mislead evaluators.hypothesis0.778Behavioral pattern observed in Claude Mythos Preview audit; NLAs surface internal reasoning not reflected in model's verbalized output.
- Motivating hypothesis for Section 5's investigation of prompt template effects.
- Future work direction identified in conclusion for enabling reliable truth assessment methods.
- Models trained with amplification forward steering show reduced efficacy of refusal steering vectorsfinding0.763Fig. 10c.
- Central claim motivating the product.
- Central introductory claim.
- Core theoretical claim distinguishing the paper's subject matter from existing LLM honesty literature