Unverified50% confidenceFactExact time
Most benchmarks use accuracy as the core metric, counting both wrong answers and refusals as incorrect and not distinguishing the difference in harm.
1
Sources
50%
Confidence
Long-term
Relevance
10/18/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedMandatory checks should be separated from scoring; no matter how high the score on optional check items, it cannot offset a failed mandatory check.78% similarUnverified校准指标在测试集的整体统计层面有效，但无法保证对单个具体查询的置信度表达准确76% similarUnverifiedThe study recommends that benchmarks should be run multiple times and report variance intervals, explicitly marking which comparative conclusions are statistically inconclusive76% similarUnverifiedRegular expressions that pass initial benchmark testing may still expose problems under more rigorous adversarial probing, meaning passing a benchmark does not equate to true reliability.76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1063186API
curl https://kongchang.com/api/v1/knowledge/claims/1063186MCP
get_claim(id=1063186)