The 25% reduction in multi-attempt rate from OTD ablation suggests additional mechanisms contribute to ESR beyond the identified latents

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

• Ablating 26 OTD latents reduces multi-attempt rate by 25% (from 7.4% to 5.5%) in Llama-3.3-70Bfinding0.854
  Primary causal evidence for dedicated internal consistency-checking circuits
• Random latent ablation produces slight increase in ESR rate (3.8% to 4.2%), not statistically significantfinding0.841
  Control result confirming OTD ablation effect is specific to those latents, not a general ablation artifact
• OTD latent ablation leaves mean first-attempt score unchanged (baseline 26.3, ablation 27.4) in Llama-3.3-70Bfinding0.787
  Evidence that OTDs specifically support meta-cognitive monitoring rather than general response generation
• The 26 differentially-activated OTD latents play a causally important role in enabling ESR in Llama-3.3-70Bclaim0.783
  Causal interpretation of the ablation experiment results
• We hypothesize ESR might be adversarially circumvented through targeted interventionshypothesis0.764
  Open safety-relevant question about whether ESR can be bypassed
• 0% multi-attempt responses across 7,892 no-steering baseline trials confirming ESR is steering-inducedfinding0.764
  Control result establishing that self-correction is specifically induced by steering, not spontaneous model behavior
• Ali et al. 2025 found contrastive activation addition less effective at larger model scale, consistent with ESR in 70Bfinding0.761
  Prior finding from related work that aligns with ESR being strongest in the largest model tested
• Five judge models agree 90-96% on multi-attempt detection and ESR direction for same responsesfinding0.735
  Validation that ESR findings are not artifacts of any particular judge model's evaluation methodology