Pairwise cosine similarities between Description, Narration, and Dialogue evil vectors are all below 0.5

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• In Qwen-2.5-9B, only v1 has meaningful cosine similarity to DIM direction; all additional basis vectors have cosine similarities ~1e-9finding0.809
  Appendix E replication of DIM alignment finding in Qwen model
• Cosine Similarity of Persona Vectorsconcept0.799
  Geometric measure used to track persona vector drift and stability across checkpoints
• Experimental condition adjective embeddings show mean cosine similarity 0.657 (n=9,591 pairs), significantly higher than history (0.628, t=15.8, p=1.4×10⁻⁵⁵), conceptual (0.587, t=38.5, p<10⁻³⁰⁰), and zero-shot (0.603, t=35.1, p=4.3×10⁻²⁶²)finding0.795
  Core result of Experiment 3: cross-model semantic convergence under self-referential processing
• Cosine similarity between truth probesmethod0.780
  Geometric evaluation of truth direction alignment across layers and prompt templates.
• Pairwise structural correlations of inter-trait cosine similarity matrices across all four architectures range from r=0.898 to r=0.986 (all p<0.001), indicating cross-architecture geometric preservation.finding0.772
  Cross-architecture geometric invariance of Big Five steering vectors
• Evil and Impolite persona vectors evolve in parallel in MDS space, suggesting intertwined representationsfinding0.768
  Geometric finding from MDS analysis suggesting shared representational structure
• Generated statements achieve 85.62%-94.00% cosine similarity alignment with Perez et al. validated OCEAN and Dark Triad statementsfinding0.768
  Validates the statement synthesis pipeline as producing behavior-specific content comparable to established methods
• Same-trait cosine similarities between GPT-4.1-mini and DeepSeek-V4-Flash extracted persona vectors are ≥0.93finding0.754
  Validates robustness of persona directions to choice of LLM judge model