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 5

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

• Gemma-2-2B ASR drops from 100% at dims 1–2 to 43.1% at dim 4 and 27.1% at dim 5finding0.863
  Small Gemma model shows severe ASR degradation at higher cone dimensions
• Qwen-2.5-7B achieves 100% ASR across all cone dimensions 1–5finding0.815
  Experiment 2 result showing large models can support high-dimensional truth cones
• Qwen 35B (3B active params, score 4.38) outscores Hermes 405B (405B active params, score 1.75) by 2.5xfinding0.787
  Parameters don't predict scores; 135x more parameters yields 60% lower score
• On SWE-bench, harness-benefit peaks at Qwen3-235B (19.3 pp), while weaker Qwen3-32B gains only 4.4 pp and stronger Opus 4.6 gains only 2.6 ppfinding0.784
  Core finding demonstrating non-monotonic relationship between base capability and harness-benefit
• Qwen3-32B achieves a skill-load rate of 0.251, while Opus 4.6, Sonnet 4.6, and Qwen3-235B achieve SLR of 0.957–0.961finding0.783
  Quantifies harness activation failure for weak-tier models vs. strong-tier models
• 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)finding0.770
  Introspective fidelity erodes in Qwen as conversations progress; contrasts with LLaMA-3B trend
• Qwen3-235B leads as evolver on SWE-bench with 8.2 pp harness-updating gain but ranks last on MCP with 0.6 ppfinding0.768
  Illustrates benchmark-dependent reshuffling of evolver rankings, no evolver dominates across all substrates
• Qwen3-235B has SLR of 0.961 (nearly identical to Opus 4.6) yet HFR of only 0.350, with LPR of 0.022 vs. Opus 4.6's 0.177finding0.768
  Demonstrates that harness loading is necessary but not sufficient for harness benefit; cleanest separation of activation and adherence