Unverified60% confidenceFactExact time
Single judge models have known biases including position bias, length bias, and self-enhancement bias.
2
Sources
60%
Confidence
Long-term
Relevance
7/10/2026
First Seen
Sources
Forget Benchmark Scores! Four Practical Standards to Judge an LLM's Real Capability
bilibili鲲鹏AI探索局7/6/2026
Related Claims
Unverified单一LLM评判者存在位置偏见（倾向于给出现在前面的答案更高评分）和自我增强偏见（倾向于给自家风格的输出打高分）84% similarVerifiedLLM-as-a-Judge方法存在位置偏差、冗长偏差和自我偏好偏差等已知系统性偏差80% similarUnverifiedJudge models exhibit systematic biases: favoring output styles similar to their own training distribution, giving higher scores to longer responses (verbosity bias), and being sensitive to the position of answers in the prompt (position bias).77% similarUnverifiedA common limitation shared by mainstream bias benchmarks is that they measure a model's explicit choices in specific controlled contexts, rather than implicit biases that emerge in open-ended generation, long-text reasoning, or complex role-playing scenarios.77% similarUnverifiedThe study locates the source of bias in a specific routing signal—input length—rather than in the overall mechanism of complexity assessment.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/412646API
curl https://kongchang.com/api/v1/knowledge/claims/412646MCP
get_claim(id=412646)