Interest introspection improves from 1B to 3B: ρ from 0.19 to 0.80, R² from 0.05 to 0.66

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

• Wellbeing introspection improves from 1B to 3B: ρ from 0.48 to 0.66, R² from 0.26 to 0.45finding0.881
  Confirms scaling trend for wellbeing concept between smallest and middle model size
• LLaMA-3.2-1B impulsivity introspection: ρ=0.21, p<10⁻⁴ (significant but weaker than 3B ρ=0.52)finding0.814
  Impulsivity shows significant introspection in 1B but declines in 8B; non-monotonic scaling
• Impulsivity→interest: ρ increases from 0.70 (α=-4) to 0.83 (α=+4); R² from 0.46 to 0.69 in LLaMA-3.2-3Bfinding0.793
  Scatter plot visualization showing strengthened probe-report relationship across alpha range
• Qwen 2.5 7B-Instruct wellbeing introspection: ρ=0.49, isotonic R²=0.76 (LMM p<10⁻¹⁰)finding0.792
  Strong introspective coupling in Qwen model; demonstrates cross-family generalization of introspective capacity
• Mean validated introspective fidelity across concept-model pairs: R²=0.12 (1B), 0.37 (3B), 0.61 (8B); pooled LMM β=0.29, p=5.55×10⁻⁹⁹finding0.791
  Strong scaling trend for introspective fidelity when excluding invalid steering-sign pairs
• Gemma 3 4B-IT wellbeing introspection: ρ=0.28, isotonic R²=0.11 (LMM p=1.33×10⁻¹³)finding0.786
  Weaker but still significant introspective coupling in Gemma model; consistent with lower probe quality
• Focus→wellbeing: ρ increases from 0.42 (α=-4) to 0.85 (α=+4); R² from 0.34 to 0.75 in LLaMA-3.2-3Bfinding0.778
  Scatter plot visualization of the dramatic tightening of probe-report relationship at extreme steering settings
• Pearson-Vogel et al.: accurate self-description prompts increase introspective detection from 0.3% to 39.9%finding0.776
  Cited to mechanistically support why the contemplative prompt changes what post-training-shaped final layers allow through