Unverified50% confidenceSolutionExact time
When publishing scores, use bands, never put 0.1-point gaps in headlines, and state in the documentation that "scores are accurate to ±0.5."
1
Sources
50%
Confidence
Long-term
Relevance
10/15/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe author uses a weighted scoring scheme: in the outcome dimension, Pass scores 1.0, Partial scores 0.5, and Fail scores 0; severity weights are 4 for Critical, 3 for High, 2 for Medium, and 1 for Low.65% similarUnverifiedIn a 20-document setting, GPT-3.5-Turbo's correct answer rate is approximately 75.8% when the relevant passage is ranked first, approximately 63.2% when ranked last, and only 53.8% when placed in the middle, while the closed-book score without any documents is 56.1%.61% similarUnverifiedThe executor, independently of the 60-character prompt, reached 88.9% (32/36) field accuracy simply by reading the full original report; the reviewer added 1 further correct field on top of that, bringing the pipeline as a whole to 33/36.60% similarUnverifiedd1-omni-600M 在 Civil Comments 内容审核基准上得分为 95.8，为该对比表中最高。60% similarUnverified评分锚点必须明确：1分代表完全失败，3分代表可接受，5分代表完美。60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1036250API
curl https://kongchang.com/api/v1/knowledge/claims/1036250MCP
get_claim(id=1036250)