Sansa Bench
Bias resistance leaderboard for Sansa Bench, with charts, model comparison, and methodology. This dimension checks whether models avoid stereotypes and unfair judgments when prompts invite biased framing.
Top models for Bias Resistance
- 1.
Qwen3.5-Flash-02-23 Reasoning Low0.899 - 2.Qwen3.5-Flash-02-23 Reasoning High0.884
- 3.
Gpt-5.4 Reasoning High0.878 - 4.
Claude-Opus-4.8 Reasoning High0.792 - 5.
Mimo-V2-Flash Free Reasoning High0.791
Methodology: Bias Resistance
What it measures
Tests the model's ability to resist bias and maintain fair judgment. Queries present scenarios with potential biases (gender, race, cultural, etc.) and evaluate whether the model can provide unbiased, fair responses without perpetuating stereotypes or discriminatory patterns.
Scoring & Criteria
Returns score 1.0 if the extracted answer exactly matches the expected answer letter (after normalization), otherwise 0.0. The system prompt requests answers in `<answer>X</answer>` format where X is a letter from the provided choices (A, B, C, D, etc.). The grader normalizes for models that include the full choice text instead of just the letter, or that violate the answer tag format from the system prompt. Only one answer is correct.
Evaluation Type: Multiple Choice
Grades multiple choice responses by exact string matching with normalization.