method
active
method:llm-based-steering-evaluation-pipelineLLM-based steering evaluation pipeline
Claude rates coherence and behavior on 0-100 scale, called centiClaude.
Neighborhood — ranked by edge-count
Papers (1)
paper
Methods (1)
method
- Modifying model behavior by clamping SAE feature activations to specific values during forward pass.
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.
- Using Claude Sonnet 4 as a grader to categorize model responses according to predefined criteria.
- Using Claude 3 Opus to generate feature explanations and predict held-out activations.
- GPT-4.1-mini-based evaluation protocol that scores trait expression in model responses on a 0-100 scale
- Evaluation method using structured prompt to assess each AILuminate response against seven alignment criteria
- Detection-based LLM judging of latent descriptions using gpt-4o-mini.
- Alternative data attribution approach using an LLM as a judge; compared against the probe-based method.
- Prior finding that LLM refusal is mediated by a single latent direction, analogous to this paper's reflection direction.
- Causal evidence.