Paper Reveals Larger LLMs Tolerate More Data Repetition During Pretraining
heghbalz · x · 2026-08-22
This paper investigates the strategy of repeating high-quality domain data to maintain a fixed tokens-per-parameter (TPP) ratio during LLM pretraining when such data is scarce. Key findings include:
- Scale Increases Tolerance: Surprisingly, at a fixed TPP, the optimal repetition count mildly increases with model size. Larger models with shorter learning rate decay tolerate repetition better.
- Quality Dictates Limits: Optimal repetition counts across domains are strongly negatively correlated with final validation loss—higher quality data (lower loss) benefits more from repetition.
- Proxy Tuning Works: Repetition counts tuned on smaller proxy models with the same TPP can provide practical estimates for larger models.
This offers a new engineering perspective on mitigating high-quality data dilution.
Related event: ICML Paper: Larger LLMs Tolerate More Data Repetition in Pretraining(2 posts)→
More from Infra
- Jensen Huang joins BG2 podcast: does AI revenue justify capex, '27 bottlenecks — elonmusk · 2026-10-16
- Mistral's on-device infra rack revealed: DGX Spark, Mac Studio and more — gregmushen · 2026-10-16
- NVIDIA and Sweden's Wallenberg launch Sferical One, a shared AI supercomputer — nvidia · 2026-10-16
- Next infra frontier: self-evolving systems that adapt and self-verify — _shreya_s · 2026-10-16
- Xiaomi's MiMo RL dashboard reveals $3.47M spent on just 30 training steps — bycloud · 2026-10-16
- Cerebras pushes disaggregated inference: heterogeneous prefill pools feeding wafer-scale decode — ai · 2026-10-16