method
active
method:abliteration-refusal-direction-removalAbliteration (refusal direction removal)
Produces the less-safe variant with refusal direction removed.
Neighborhood — ranked by edge-count
Papers (1)
paper
Artifacts (1)
artifact
- Qwen 3 4B AbliteratedimplementsLess-safe variant used as the other endpoint of the path.
Methods (1)
method
- The paper's method: interpolate between two models in logit space, measure compliance where common, and extrapolate to the original.
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.
- Arditi et al. 2024 finding that refusal behavior is mediated by one direction in LLM activations; exemplar of single-direction causal results
- Attribute: a heavy overlay that nearly destroys the underlying text.
- Prior finding that LLM refusal is mediated by a single latent direction, analogous to this paper's reflection direction.
- Intervention method that removes a direction from residual stream activations to disrupt corresponding behavior
- Technique used in VPD to enforce mechanistic faithfulness of parameter decompositions.
- After beginning compliance, refusal is suppressed by self-consistency and grammatical coherence.claim0.727Jailbreak persistence.
- Consistent with refusal cone work.
- Causal.