Typicality bias rate exceeds 50% chance baseline by 4-12 percentage points across all base models on all four preference datasets

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

• Typicality Bias Ratemethod0.853
  Measurement of how often human annotators prefer the response with higher base model log-probability
• Typicality bias rates in instruction-tuned models remain at similar or higher levels compared to their base model counterpartsfinding0.841
  Shows typicality bias is preserved through instruction tuning and RLHF, not introduced by alignment
• Typicality bias is present across all annotator demographic groups but varies in magnitude, potentially reflecting differences in linguistic norms captured by pretraining dataclaim0.831
  Finding from PRISM dataset analysis showing typicality bias varies by ethnicity and region, with implications for fairness
• Typicality Biasconcept0.813
  The human tendency to prefer more typical, familiar, fluent, and predictable text in annotation tasks, identified as a fundamental data-level cause of mode collapse
• Typicality bias in preference data is a fundamental and pervasive data-level cause of mode collapse in aligned LLMsclaim0.812
  The central thesis of the paper, distinguishing it from algorithmic explanations of mode collapse
• Typicality bias with weight α>0 mathematically sharpens the aligned policy via a power transform of the reference distributionclaim0.796
  Theoretical result showing that any positive typicality bias weight γ-sharpens the reference distribution, amplifying modes
• Varying baseline cutoff over {65,70,75} changes no S/N/I labels in either model; gain cutoff {5,10,15} moves at most 10 traits per modelfinding0.772
  Demonstrates robustness of the trichotomy classification to cutoff choice
• All VS variants maintain refusal rates above 97% on StrongReject benchmark, within 0.3-0.8 percentage points of the Direct baselinefinding0.758
  Confirms VS does not compromise safety alignment while improving diversity