method
active
method:pca-on-persona-spacePCA on Persona Space
Standardized PCA run on role vectors to find main axes of persona variation
Neighborhood — ranked by edge-count
Methods (1)
method
- Pipeline for extracting mean post-MLP residual stream activations from model responses under persona-specific system prompts to produce role vectors
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.
- Method used by Lu et al. to find orthogonal directions of maximum variance among 275 character archetypes in activation space
- Low-dimensional space of activation directions corresponding to diverse character archetypes in LLMs
- Finding establishing cross-model consistency of the assistant axis as the dominant structure in persona space
- Dimensionality reduction applied to residual stream embeddings to visualize cyclic fixed point trajectories
- Statistical method used to analyze neural activity data.
- Used to visually inspect separation of truth-related directions in model activation space across layers
- An SAE latent representing a simulated character with consistent traits, as opposed to a specific trait
- Stable regions or basins of attraction in persona space that correspond to coherent dispositional profiles