method
active
method:contrastive-prompt-persona-feature-pipelineContrastive-prompt persona feature pipeline
Rank features by activation ratio on positive vs negative rollouts.
Neighborhood — ranked by edge-count
Papers (1)
paper
- Turn-Averaged SAEs for Feature Discovery and Long-Context Attributionintroducesmentions
Methods (1)
method
- Persona Vectorsanalogous_toDirections in activation space encoding a model's dispositional self-presentation, used as evidence for self-modelling.
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.
- An SAE latent representing a simulated character with consistent traits, as opposed to a specific trait
- Technique for extracting trait directions by contrasting model activations under trait-eliciting vs. trait-suppressing conditions
- Named procedure for extracting persona vectors from mean residual-stream activation differences between trait-expressing and non-expressing responses
- A pipeline employing controlled semantic oppositions to distill monosemantic functional features from sparse activation spaces.
- Adaptation of instruction-tuned extraction to base models using third-person descriptions and hypothetical situations
- Training method for probes: generate completions under opposing system prompts to induce positive and negative poles of a concept
- Low max cosine similarities.
- Method of extracting persona vectors by contrasting activations when model is prompted to exhibit vs suppress a trait