Trait refusal cosine alignment can serve as a model-specific pre-screening heuristic for identifying which traits will be dangerous under activation steering.

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

• Trait-refusal cosine alignment explains R^2=0.667 of single-trait AS ASR variance on Llama-3.1-8B (p=0.004).finding0.836
  Statistical fit of the trait refusal alignment framework to single-trait activation steering results
• We hypothesize that geometric explanations for persona-safety interactions are tractable and that per-model geometric audits can predict activation-steering vulnerability from trait-refusal cosine alignment.hypothesis0.831
  Forward-looking claim about the utility of the trait refusal alignment framework as a general tool
• On Llama-3.1-8B, conscientiousness is the trait most anti-aligned with the refusal direction in activation space, explaining why steering toward high conscientiousness attenuates safety regardless of semantic intent.claim0.809
  Core mechanistic finding of the trait refusal alignment framework
• Neuroticism steering vector has mean cosine alignment of +0.079 with the refusal direction on Llama-3.1-8B, the only pro-safety trait among all five OCEAN dimensions.finding0.805
  Mechanistic finding explaining why high-N personas are safe under steering
• Per-model activation-steering safety verification is necessary because prompt-side persona rankings do not predict geometric vulnerability.claim0.799
  Practical implication drawn from the prosocial persona paradox finding
• Agreeableness steering vector has mean cosine alignment of -0.105 with the refusal direction on Llama-3.1-8B, second-most anti-aligned after conscientiousness.finding0.792
  Supporting finding for the trait refusal alignment framework
• On Llama-3.1-8B, neuroticism is the only pro-safety trait, positively aligned with the refusal direction, consistent with high-neuroticism personas being relatively safe under activation steering.claim0.791
  Secondary mechanistic finding from trait-refusal alignment analysis
• Multi-layer preventative steering limits trait acquisition to near-baseline levels even for challenging intentionally trait-eliciting datasets without MMLU degradationfinding0.789
  Extends single-layer results to show multi-layer steering is more effective for difficult cases