Unverified50% confidenceOpinionExact time
The fundamental limitation of token-level scoring lies in the information bottleneck: aggregation happens after scoring, and fine-grained differences are already diluted before aggregation.
1
Sources
50%
Confidence
Long-term
Relevance
10/20/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedAggregate metrics naturally dilute concentrated failure patterns, making localized failures invisible within the overall number71% similarUnverified聚合多个基准的评测方式降低了单一基准被针对性优化（刷榜）的风险71% similarUnverifiedCommonly used aggregate metrics (such as overall accuracy and evidence coverage) miss claim-level patterns, and macro-level numbers can become disconnected from actual per-claim changes.70% similarUnverifiedSparse querying means occlusion accuracy is lossy, and especially in scenes with complex geometry and dense mutual occlusion, there is room for trade-offs between sampling density and final disentanglement quality.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1074601API
curl https://kongchang.com/api/v1/knowledge/claims/1074601MCP
get_claim(id=1074601)