Unverified50% confidenceBenchmarkExact time
在冻结重放测试场景中,为了实现完整召回,部分案例需要读取407至454个候选中的290到426个
1
Sources
50%
Confidence
Long-term
Relevance
9/15/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedDiscoBench最终生成211个测试样本,包含463个歧义实例64% similarUnverified在综合基准测试中,So得55分(约78.6%),Terra得44分(约62.9%),Luna得31分(约44.3%)56% similarUnverified本次测试共产生147次模型调用55% similarUnverified在512-4096上下文长度测试中,将草稿长度从4字符拉满至16字符,整体推理延迟仅增加0.2%-1.3%52% similarUnverified叠加cv=5的5折交叉验证后,144种参数组合需训练144×5=720次模型51% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/922247API
curl https://kongchang.com/api/v1/knowledge/claims/922247MCP
get_claim(id=922247)