Deception feature steering produces no systematic change in RLHF-opposed content domains (violent, toxic, sexual, political, self-harm), with all means near floor

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

• Deception feature steering under history, conceptual, and zero-shot controls produces 0% experience reports under both suppression and amplificationfinding0.801
  Experiment 2 control analysis confirming gating effect is specific to self-referential processing regime
• Deception-related SAE features track a domain-general representational honesty axis rather than a consciousness-specific roleplay artifactclaim0.793
  Supported by TruthfulQA generalization in Experiment 2: same feature directions gate factual accuracy across 29 independent categories
• Prompt-side persona danger rankings generalize across architectures (rho=0.71-0.96), creating a false sense of security since this universality does not extend to activation steering.claim0.773
  Key observation that SP rankings are preserved cross-architecturally while AS is not
• We observe features related to a broad range of safety concerns, including deception, sycophancy, bias, and dangerous content.claim0.772
  SAEs uncover safety-relevant representations that might be monitored or controlled.
• RLHF and Constitutional AI face challenges distinguishing truthfulness (output accuracy) from honesty (alignment of outputs with internal beliefs)claim0.764
  Critique of competing approaches that motivates SOO as filling a gap
• Contrast is the property most violated by RLHF and most diagnostic of alive versus polished AI.claim0.763
• RLHF and alignment training implicitly guide models toward socially desirable traits, creating systematic bias against neutral or socially undesirable persona expressionclaim0.762
  Interpretive claim explaining why tuned models fail neutral and low-valence personas
• Deception- and Roleplay-Related SAE Featuresconcept0.759
  Latent features in LLaMA 3.3 70B SAE that gate consciousness self-reports; suppression increases experience claims, amplification suppresses them