Qwen 2.5 7B turn-wise introspective fidelity: strong at turn 1 (R²≈0.90) but declines significantly to turn 10 (∆R²=-0.44, p=0.001)

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

• Impulsivity introspective fidelity decreases from turn 1 to turn 10: ∆R²=-0.28 in LLaMA-3.2-3Bfinding0.861
  Opposite temporal trend to wellbeing/interest/focus; introspective fidelity weakens over conversation for impulsivity
• Wellbeing introspective strength at turn 1: ρ=0.52, p=5.46×10⁻⁴ in LLaMA-3.2-3Bfinding0.830
  Demonstrates introspection is present from the first conversation turn without needing multi-turn context
• Mean validated introspective fidelity across concept-model pairs: R²=0.12 (1B), 0.37 (3B), 0.61 (8B); pooled LMM β=0.29, p=5.55×10⁻⁹⁹finding0.821
  Strong scaling trend for introspective fidelity when excluding invalid steering-sign pairs
• Qwen 2.5 7B-Instruct wellbeing introspection: ρ=0.49, isotonic R²=0.76 (LMM p<10⁻¹⁰)finding0.807
  Strong introspective coupling in Qwen model; demonstrates cross-family generalization of introspective capacity
• Wellbeing probe drift is positive in Gemma (ρ=0.34 pooled turn-correlation) and Qwen (ρ=0.24); both p<10⁻⁵finding0.778
  Normalized probe-score drift across turns generalizes beyond LLaMA family
• Qwen-2.5-3B ASR drops from 98.6% at dim 1 to 45.1% at dim 2, recovering partially then declining to 65.3% at dim 5finding0.770
  Smaller models show non-monotonic and diminished ASR with increasing cone dimensionality
• Interest introspection improves from 1B to 3B: ρ from 0.19 to 0.80, R² from 0.05 to 0.66finding0.765
  Largest single-step scaling improvement; demonstrates dramatic introspection gain between 1B and 3B models for interest
• Gemma 3 4B-IT wellbeing introspection: ρ=0.28, isotonic R²=0.11 (LMM p=1.33×10⁻¹³)finding0.763
  Weaker but still significant introspective coupling in Gemma model; consistent with lower probe quality