Unverified60% confidenceFactExact time
The training-time scaling law was proposed by Kaplan et al. in 2020, holding that model capability is primarily determined by parameter count, training data volume, and training compute.
2
Sources
60%
Confidence
Long-term
Relevance
7/8/2026
First Seen
Sources
Verification Loops Quadruple DeepSeek's Intelligence at One-Seventh the Cost of Claude Opus
hackernewshackernews7/7/2026
Related Claims
UnverifiedKaplan等人2020年的研究表明模型参数量、训练数据量、训练计算量三者与性能之间存在可预测的幂律关系82% similarUnverifiedTraditional Scaling Laws describe the power-law relationship between model capability and training parameter count, data volume, and compute, with the implicit premise that a model's capability is fixed once it is trained.76% similarUnverifiedThe academic community is advocating a "compute threshold" regulatory framework that mandates reporting for models trained with compute exceeding a certain magnitude.72% similarVerifiedThe Scaling Law proposed by OpenAI reveals that model performance improves in a predictable power-law fashion as the number of parameters, training data volume, and compute increase.71% similarUnverified美国AI监管以模型训练算力阈值（以FLOP为单位）作为触发强制报备义务的量化指标71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/274900API
curl https://kongchang.com/api/v1/knowledge/claims/274900MCP
get_claim(id=274900)