Unverified80% confidenceFactTime unknown
K2-Vendor-Verifier通过logprobs、token级别一致性等指标来量化模型精度差异
1
Sources
80%
Confidence
Long-term
Relevance
6/1/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedK2-Vendor-Verifier使用固定输入测试集覆盖代码生成、推理、对话等多种任务类型进行验证66% similarUnverifiedRAG system recall evaluation should use quantitative metrics such as Recall@K, MRR, and NDCG for closed-loop assessment.57% similarUnverifiedIn k-fold cross-validation, each fold produces a set of performance metric values, providing paired observation samples for the Wilcoxon signed-rank test.57% similarUnverifiedE2AM indicates whether each value is measured or estimated, and falls back to a power limit × utilization estimation method if the GPU does not expose NVML power sensors.56% similarUnverified校准性指模型表达的置信度与实际准确率之间的一致程度55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/29548API
curl https://kongchang.com/api/v1/knowledge/claims/29548MCP
get_claim(id=29548)