Unverified50% confidenceFactExact time
In the extended trials on 26 models, numerical computation performance rises together with non-numerical performance.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/20/2026
First Seen
Valid until: 1/18/2027
Sources
Related Entities
Related Claims
Unverified测试时计算扩展的边际收益并非线性，存在明显的收益递减区间71% similarUnverifiedFor small models with limited parameter counts, data quality can have an even greater impact on final performance than data quantity.68% similarUnverifiedIn the ablation study, when no data is used and only parametric memory is relied on, even the best model performs poorly; after providing early-quarter data, performance improves by about 25 percentage points; but there is still a 12-16 percentage point gap from using the latest data.66% similarUnverifiedThe power-law exponent between test loss and compute is approximately 0.05, meaning that every 10x increase in compute yields approximately a 12% improvement in model performance.66% similarUnverified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66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1073472API
curl https://kongchang.com/api/v1/knowledge/claims/1073472MCP
get_claim(id=1073472)