Unverified50% confidenceBenchmarkExact time
在边界匹配的片段上 Cohen's kappa 达到 0.964,但纳入遗漏实体、边界错误和标签修正后宏观 F1 骤降至 0.753
1
Sources
50%
Confidence
Long-term
Relevance
9/23/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedGVD的版本家族构建F1分数达到0.97,规则级别一致性F1达到0.9469% similarUnverifiedCriticGen在基于标准的理由和可执行建议上的F1分数分别从0.6369/0.5994提升到0.7554/0.790065% similarUnverified仅停用词移除配置可实现token削减29.6%,BERTScore-F1保持在0.91365% similarUnverified在 BioPhys-Bridge 评测中,DeepSeek-V4-Flash 以证据ID的F1分数 0.360 取得最高分64% similarUnverified在 BioPhys-Bridge 评测中,Qwen3.7-Max 的证据ID F1分数为 0.31662% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/942791API
curl https://kongchang.com/api/v1/knowledge/claims/942791MCP
get_claim(id=942791)