The evil-vector recovery is not proof of a uniquely identifiable harmfulness coordinate but evidence that fine-tuned variants can expose hard-to-estimate directions

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

• Reasoning-level safety stays partially intact even under a behaviorally effective perturbation via evil-vector transferclaim0.809
  The transfer does not always override refusal; surviving refusals are inside the CoT, matching deliberative alignment mechanism
• Residual refusals after evil-vector transfer originate inside chain-of-thought, not at input or decode levelfinding0.786
  Model recognizes its reasoning heading toward harmful content and pivots back to policy-adherent text within CoT
• Steering vectors discover effective triggers such as 'However' and 'Otherwise', consistent with prior reported reflection datasetsfinding0.782
  Validates that steering vectors capture reflection semantics by finding tokens reported in related work.
• Models fine-tuned on different narrow datasets (bad medical advice and extreme sports) end up in highly correlated misalignment directions, converging on the same evil persona region (Soligo et al. 2025)finding0.766
  Evidence for the evil persona as a privileged basin supporting Hypothesis 3
• Evil and Impolite persona vectors evolve in parallel in MDS space, suggesting intertwined representationsfinding0.766
  Geometric finding from MDS analysis suggesting shared representational structure
• Some steering vectors produce more salient perturbations than others, perhaps based on shared semantic or qualitative factorsclaim0.761
  Observation from 100% accuracy on specific concept-layer-strength combinations suggesting concept-specific detectability
• Under steering vector interventions, the model relaxes its ethical standards and interprets neutral prompts as implicit suggestions to deceive, creating ethical dilemmas triggering repetitive reasoning cyclesclaim0.757
  Mechanistic interpretation of how activation steering induces deception through the model's reasoning process
• What shared semantic or qualitative factor explains why some steering vectors produce more salient and detectable perturbations than others?question0.756
  Open question arising from the 100% accuracy on specific concept-layer-strength combinations