claim
active
claim:activation-steering-works-better-when-explaining-away-on-rejected-responses-than-amplifying-on-chosen-responsesActivation steering works better when explaining away on rejected responses than amplifying on chosen responses
Model may fight chosen-response steering.
Source paper
extracted_from(2026) · Leon Bergen · Usha Bhalla · Sidharth Baskaran · Max Loeffler +13
Neighborhood — ranked by edge-count
Papers (1)
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.
- Nuanced interpretive claim about the limits of steering as a mechanism for reflection enhancement.
- Core validation that identified latent directions correspond to meaningful control over reflective behavior.
- Steering derived from LRH.
- Central claim of the paper; supported by the model organism ground-truth approach.
- Coherence advantage claim with mechanistic speculation about why steering leads to incoherence
- Key distinction showing steering offers value beyond prompting; supported by Figure 5 and random vector experiments.
- Key intervention result showing steering vectors can induce deceptive behavior from a neutral baseline
- Applied security implication derived from the asymmetry finding.