concept
active
concept:effective-skill-unlearning-through-intervention-and-abstention-li-et-al-2025Effective skill unlearning through intervention and abstention (Li et al., 2025)
Related work achieving machine unlearning via activation steering
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.
- Conditional negative clamping on WMDP-bio with MMLU retention.
- Removing targeted knowledge from models.
- Training that suppresses representations of undesirable concepts may incentivize new representations, removing the detection signalhypothesis0.745Tension for monitorability.
- What training interventions could reliably prevent alignment faking without compromising model capabilities?question0.735Practical implication of the paper's findings; not resolved