Refusal Vector

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Deception Vectorconcept0.778
  Extracted steering vector capturing semantic dimension of strategic deception in moral dilemmas in Experiment 1
• Refusal directionconcept0.770
  Arditi et al. 2024 finding that refusal behavior is mediated by one direction in LLM activations; exemplar of single-direction causal results
• refusal rateconcept0.752
  The percentage of harmful requests that a model refuses to answer, a common safety metric.
• Sense Vectorsconcept0.737
  Vectors acquired during pretraining in Backpack LMs that have a multiplication effect on model generation
• Persona Vectors (Chen et al.)framework0.730
  Prior framework for monitoring and controlling character traits in LLMs via activation directions; this paper extends it to 275 roles
• steering vectorsconcept0.725
  A method for modifying model behavior by adding perturbation vectors to activations, used here to try to reduce eval awareness.
• Refusal Direction in LLMsconcept0.716
  Prior finding that LLM refusal is mediated by a single latent direction, analogous to this paper's reflection direction.
• Random vector baselinemethod0.716
  Baseline method sampling a random vector as feature direction for comparison with learned methods