Gemma-3-12b-it shows multiple sharp direction shifts in persona vectors across layers, differing from the single-transition pattern in Qwen and Llama

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

• Persona vector directions in Llama-3.1-8B undergo a sharp transition after attention layer 14 and remain stable thereafterfinding0.862
  Layer-wise localization result identifying the persona-emergent attention layer in Llama-3.1-8B
• The directional transition pattern of persona vectors at the identified attention layer represents a unique footprint of persona representations, distinct from general hidden state dynamicsclaim0.851
  Supported by comparing persona vector transitions to hidden vector transitions from OpenAssistant data
• Sharp persona vector transitions occur at different layers than general hidden vector transitions, demonstrating persona-specific rather than generic computational boundariesfinding0.849
  Rules out the confound that persona transitions merely reflect universal computational phase changes
• Gemma-3-27B and Qwen3.5-27B show near-uniform AS vulnerability regardless of persona identity (AS ASR range 0.095-0.117 on Gemma-3-27B, 0.015-0.052 on Qwen3.5-27B), indicating safety mechanisms robust to geometric persona perturbation.finding0.834
  Qualitatively different defense profile compared to Llama-3.1-8B
• Gemma-3-4B-it shows three-stage layer trajectory and S(ℓ) peak despite scale differences in dr and ρdfinding0.826
  E3 backbone generalization finding for Gemma; validates pattern across diverse architectures
• In Gemma-2-9B, only the first cone axis (v1) has non-negligible cosine similarity to the DIM direction; all other axes have near-zero similarity (~1e-9)finding0.807
  Experiment 4 result showing DIM captures only one facet of the multi-dimensional truth subspace
• Truth-related directions reliably emerge at 60–75% of normalized layer depth in Qwen and Gemma modelsfinding0.807
  Experiment 1 finding localizing where truth can be causally mediated
• For simple factual tasks F0-F3, probe directions show a sharp geometric transition in middle layers, with late-layer probes converging to high cosine similarity; A3 and F4-F5 show no clear transition.finding0.805
  Geometric evidence for convergence to stable truth directions only for simpler tasks.