method
active
method:pairwise-cosine-similarity-analysisPairwise Cosine Similarity Analysis
Used to quantify the semantic clustering of adjective-set embeddings across model families and conditions
Neighborhood — ranked by edge-count
Concepts (1)
concept
- Cross-Model Semantic ConvergenceaboutimplementsThe tighter clustering of experience-report embeddings across independently trained model families under self-referential processing
Methods (3)
method
- Layer-wise Cosine Similarity Analysisrelated_toGeometric analysis tracking how persona vector directions evolve across transformer layers to identify the transition layer
- Cosine Similarity Measurementrelated_toUsed to measure alignment between DIM direction and cone basis vectors to assess overlap
- Embedding model used to compute vector representations of adjective sets for cosine similarity analysis in Experiment 3
Artifacts (1)
artifact
- Key paper finding structured first-person descriptions in LLMs claiming awareness or subjective experience during self-referential processing.
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.
- Detects tiling in real SAEs.
- Classifier using cosine similarity between activation vectors and steering vectors to detect deception with 89% accuracy
- Method to discover new reflection-inducing instructions by ranking candidate tokens by cosine similarity to steering vectors.
- Measures QK geometry between country queries and city keys.
- Geometric evaluation of truth direction alignment across layers and prompt templates.
- Cosine similarity between feature activations restricted to tokens where one of the features fires; used to identify feature splitting relationships
- Detection mechanism computing cosine similarity between activation vectors and steering vectors to classify deception
- Inspiration for distance-matrix comparison.