Unverified50% confidenceBenchmarkExact time
After correcting the scorer, all four models achieved 100% on all five Sanity gate tasks.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/11/2026
First Seen
Valid until: 10/25/2026
Sources
Related Entities
Related Claims
Unverified当所有模型在某评测上都能拿90分时该评测就应退场，即基准饱和现象66% similarUnverifiedOn CWRU, the ensemble model achieved 100% accuracy on three of the four held-out loads, and 92.27% on the fourth.65% similarUnverifiedOn action appropriateness, the average score across all models reached 95.6%.64% similarVerifiedThe chain-of-thought effect is particularly pronounced in models with parameter counts exceeding 100B.64% similarUnverified在5%显著性水平下测试100个随机因子，期望会有5个看起来显著（多重假设检验问题）64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/999500API
curl https://kongchang.com/api/v1/knowledge/claims/999500MCP
get_claim(id=999500)