Optimal activation capping layers for Qwen 3 32B are layers 46-53 (out of 64) at 25th percentile cap

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

• Optimal activation capping layers for Llama 3.3 70B are layers 56-71 (out of 80) at 25th percentile capfinding0.874
  Specific implementation finding for Llama capping parameters
• Qwen3.5-27B shows mean AS ASR of only 0.035 [0.03, 0.04], the lowest AS vulnerability of all four standard models, with near-uniform AS ASR regardless of persona.finding0.766
  Vulnerability profile for Qwen3.5-27B showing near-zero AS vulnerability
• 25th percentile of Assistant Axis projection distribution gives the most Pareto-optimal safety-capability tradeoff for activation capping, and approximately matches mean Assistant response activationfinding0.757
  Calibration finding for choosing the activation cap threshold
• Injecting CVs at mid-residual layers with the strongest stylistic effects (layer 15 for Qwen, layer 16 for Mistral) yields optimal performanceclaim0.754
  Design choice claim based on prior work and preliminary results
• Model activation along the assistant axis drifts steadily away from the assistant pole as conversation shifts toward Aura behavior in Qwen 3 32B, and activation capping eliminates Aura behaviorfinding0.750
  Finding confirming that the Aura persona shift is real, trackable, and causally relevant in persona space
• Some activation capping settings slightly improve performance on IFEval, MMLU Pro, or GSM8k for both Qwen and Llamafinding0.746
  Unexpected positive finding suggesting capping may sometimes help capabilities
• The case at approximately the 2/3 layer of LLaMA3.1-8B (Layer 24, satisfying Criteria 1 and 2) aligns with prior studies showing the 2/3 layer optimally predicts human brain activity.finding0.743
  Connects this study's results to Schrimpf et al. 2021 and Caucheteux et al. 2022/2023 findings on brain-LLM alignment.
• Meta-LLaMA-3.1-8B-Instruct shows optimal anchoring at layer 9 (S ≈ −1.90, median peak layer ℓ* = 10 [IQR 0.384])finding0.743
  E3 result establishing the Goldilocks zone at mid-layers for LLaMA architecture