Unverified50% confidenceFactExact time
The same evidence can improve the performance of weaker models while degrading the performance of stronger models.
1
Sources
50%
Confidence
Long-term
Relevance
9/8/2026
First Seen
Sources
Related Claims
Unverified小模型对数据分布更敏感，同样的数据补充量在4B模型上的边际效果往往比大模型更明显71% similarUnverifiedModels with more parameters generally have stronger comprehension capabilities, but respond more slowly and consume more tokens.71% similarVerifiedResearch shows that the output quality of the same model can vary by several times under different prompts.70% similarUnverifiedFor small models with limited parameter counts, data quality can have an even greater impact on final performance than data quantity.67% similarUnverified研究表明即使是部分污染（模型只见过题目而非答案），也可能显著提升得分67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/875704API
curl https://kongchang.com/api/v1/knowledge/claims/875704MCP
get_claim(id=875704)