method
active
method:emotion-vector-extraction-from-synthetic-storiesEmotion vector extraction from synthetic stories
Average residual activations across stories per emotion, subtract cross-emotion mean, project out neutral PCs.
Neighborhood — ranked by edge-count
Papers (1)
paper
- Emotion Concepts and their Function in a Large Language Modelintroducesmentions
Concepts (1)
concept
- Emotion vectorsimplementsLinear representations of emotion concepts extracted from activations.
Datasets (1)
dataset
- 171 emotions x 100 topics x 12 stories generated by Sonnet 4.5.
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.
- Replicates Lu et al. / Chen et al. contrastive-pair recipe.
- Method using activations from the prompt 'Tell me about {word}' minus mean over other random words to obtain concept vectors.
- The paper's core automated pipeline that takes a trait name and description as input and outputs a corresponding persona vector via contrastive prompting
- Method of extracting persona vectors by contrasting activations when model is prompted to exhibit vs suppress a trait
- Pipeline for extracting mean post-MLP residual stream activations from model responses under persona-specific system prompts to produce role vectors
- Method for obtaining concept vectors by subtracting activations from two contrasting prompts.
- Sanity check, not surprising.
- Core method for extracting persona vectors by contrasting mean activations under persona-eliciting vs. suppressing prompts