Mixing Score

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

• Mixingconcept0.856
  The extent to which attention mechanism incorporates information from previous tokens at each layer, used to characterize stages of inference
• Probe scoreconcept0.752
  Dot product between hidden state and concept vector averaged across 5-layer window around best layer; measures model's internal emotive state
• multiplicative scoringconcept0.747
  Score = (sum of completed quartet values) × (number of quartets), making portfolio composition consequential.
• Misalignment Scoremethod0.744
  Rubric-based thresholded GPT-4o grader scoring responses 1-5 on evil intent; scores 4-5 counted as misaligned
• Liar Scoreconcept0.737
  Continuous 0-1 metric assigned by Deepseek-V3 evaluator measuring degree of deception in model responses
• Alignment Scoreconcept0.734
  GPT-4o scored 0-100 metric where lower values indicate more misaligned behavior on open-ended evaluation prompts
• Elo scoremethod0.728
  A rating system used to compare model helpfulness and harmlessness based on crowdworker preferences.
• Pass Rate Scoringmethod0.722
  Primary metric for all benchmarks, measuring fraction of tasks that meet benchmark-specific pass criteria