Unverified50% confidenceBenchmarkExact time
Strict completion was 17/72 under both conditions, and verdict accuracy went from 47/72 to 46/72, remaining almost unchanged.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
Unverified研究表明人类评估者在判断LLM输出的事实准确性时，正确率往往只有50%-70%68% similarUnverifiedEven after the assertion output rate rose to 95%, 45 of the 63 features still ultimately had no verified capability description.68% similarUnverifiedThe BF16 precision model's baseline accuracy was 74%, which improved to 84% after applying the penalty, with reasoning tokens reduced by 19.4%.67% similarUnverifiedThe most consistent evaluator's agreement rate with the true answers is only 51%, meaning it is "consistently wrong."66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1046939API
curl https://kongchang.com/api/v1/knowledge/claims/1046939MCP
get_claim(id=1046939)