Larger models should amplify bias less than smaller models, with model biases more accurately reflecting data biases rather than exacerbating them

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

• The model tends to reflect more when the question is difficult, and accuracy is generally lower for harder questionshypothesis0.823
  Hypothesis explaining negative correlation between reflection rate and accuracy without implying reflection is harmful
• Bigger models are more likely to converge to a shared representation than smaller modelshypothesis0.809
  Selective pressure toward convergence via model capacity
• Typicality bias rates in instruction-tuned models remain at similar or higher levels compared to their base model counterpartsfinding0.796
  Shows typicality bias is preserved through instruction tuning and RLHF, not introduced by alignment
• Scaling may reduce hallucination and certain kinds of bias as models converge toward an accurate model of realityclaim0.793
  Implication of PRH: larger models should amplify bias less and hallucinate less if they better model reality
• Models with as few as 14B parameters can exhibit 40% misalignment while maintaining 99% coherence (Turner et al. 2025 result)finding0.788
  Concurrent work result showing emergent misalignment occurs in small models
• Using more than two models in a MAS comparison could harm alignment due to conflicting loss gradients, or could assist in isolating causal subspaceshypothesis0.775
  Open question raised in the paper about scaling MAS beyond two models.
• Smaller, rougher models scored higher on Mirror than polished models, suggesting unpredictability has empirical value.claim0.774
• Models fine-tuned on human data show strong positive correlation between misalignment and incoherence; misalignment scores fall below 1% when incoherence threshold is appliedfinding0.773
  Human data fine-tuning effect is distinct from synthetic emergent misalignment and likely caused by off-policy training