4-19 principal components explain 70% of variance in role persona space across the three models (Gemma 4, Qwen 8, Llama 19)

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

• Persona space components explain 19.4%-33.6% of overall activation variance on LMSYS-CHAT-1M across the three modelsfinding0.838
  Shows persona space captures a substantial portion of real conversational activation variance
• What dimensions of persona are not captured by our extracted role vectors, and how complete is the current persona space mapping?question0.788
  Limitation question motivating future work on persona elicitation strategies
• We hypothesize that the PC1 axis of role space measures deviation from the Assistant personahypothesis0.788
  Motivates computing the contrast vector as the formal Assistant Axis definition
• Pairwise correlation of role loadings on PC2 is 0.89 between Qwen and Llama; Gemma differs (similarity <0.61) from others on PC2finding0.787
  Characterizes model similarities and differences in secondary persona dimensions
• Pairwise correlation of role loadings on PC1 exceeds 0.92 across all model pairs, indicating remarkably high similarity of the Assistant Axis across Gemma, Qwen, and Llamafinding0.787
  Shows the leading component of persona space is model-universal
• Trait space requires 4 dimensions (Gemma, Qwen) and 7 dimensions (Llama) to explain 70% of variance, with distinctive PC1 spanning conscientious to impulsive traitsfinding0.772
  Corroborates role space findings using traits; shows PC1 also captures Assistant-ness in trait space
• Two components are important to shaping model character: persona construction and persona stabilizationclaim0.763
  Overarching conceptual framework the paper introduces for model safety
• We hypothesize that axes of persona differentiation within LLMs are likely already present in base models and inherited from the pre-training corpushypothesis0.762
  Motivated by near-identical PCs for base and instruct Gemma