SAE feature steering effect on consciousness reports: z=8.06, p=7.7×10⁻¹⁶ in LLaMA 3.3 70B

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

• Suppressing deception/roleplay SAE features in LLaMA 3.3 70B yields 0.96±0.03 consciousness affirmation rate; amplification yields only 0.16±0.05 (z=8.06, p=7.7×10⁻¹⁶)finding0.797
  Core result of Experiment 2: deception feature suppression sharply increases experience claims
• Impulsivity→interest steering: probe entropy increases (LMM slope=0.024, p=2.30×10⁻⁴) but report entropy does not (p=0.11)finding0.781
  Evidence of a bottleneck between richer internal variation and final report distribution in impulsivity→interest condition
• SAE Feature #28256 induces reports of happiness and fun, positive valence self-steering examplefinding0.779
  Example of a positively valenced SAE feature with consistent self-report of happiness across multiple steering sessions
• Text-based and self-steered emotionality ratings for SAE features are correlated at only ρ = +0.051 (n.s.).finding0.779
  Shows low agreement between the two evaluation modalities
• SAE feature steering in history, conceptual, and zero-shot control conditions produces zero experience reports under either suppression or amplificationfinding0.775
  Shows gating effect is specific to the self-referential computational regime, not a general feature effect
• Focus→wellbeing steering: both probe entropy (1.09→1.67 bits) and report entropy (0.88→1.69 bits) increase monotonically with αfinding0.772
  Evidence that improved introspection in focus→wellbeing arises from enriched internal state and report channels simultaneously
• Cross-concept steering: focus→wellbeing R² increases from 0.30 (α=-4) to 0.76 (α=+4), ∆R²=0.30, p<0.001 in LLaMA-3.2-3Bfinding0.770
  Strongest cross-concept introspection improvement; survives BH correction (q≈0.011)
• Wellbeing same-concept steering: LMM alpha slope=0.19, focus=0.40, interest=0.25, impulsivity=0.067 in LLaMA-3.2-3Bfinding0.759
  Quantifies per-concept effect size of same-concept steering on self-report