Unverified50% confidenceTradeoffExact time
The gains from increasing model parameters are approximately twice those from performing a second pass over the data.
1
Sources
50%
Confidence
Long-term
Relevance
10/9/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn model selection, pursuing the last few percentage points of performance often requires paying several times or even dozens of times more in cost.72% similarUnverified对同一测试用例执行多次调用统计成功比例得到通过率，比二元的通过/失败判断更能反映模型真实表现69% similarUnverifiedEvaluations such as LLM-as-a-Judge incur double or even multiple times the token overhead, because tokens are consumed both to generate the answer being evaluated and to have another model score it.67% similarUnverified研究表明主动检索的学习效果是被动复习的两倍以上，这一现象被称为「测试效应」（Testing Effect）67% similarUnverifiedTwo models with similar intelligence scores may differ in inference cost by several times or even an order of magnitude.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/990277API
curl https://kongchang.com/api/v1/knowledge/claims/990277MCP
get_claim(id=990277)