Not All Tokens Are What You Need for Pretraining

Abstract

Previous language model pre-training methods have uniformly applied a next-token prediction loss to all training tokens. Challenging this norm, we posit that “Not all tokens in a corpus are equally important for language model training” . Our initial analysis examines token-level training dynamics of language model, revealing distinct loss patterns for different tokens. Leveraging these insights, we introduce a new language model called R HO -1. Unlike traditional LMs that learn to predict every next token in a corpus, R HO -1 employs Selective Language Modeling (SLM), which selectively trains on useful tokens that aligned with the desired distribution. This approach involves scoring tokens using a reference model, and then training the language model with a focused loss on tokens with higher scores. When continual pretraining on 15B OpenWebMath corpus, R HO -1 yields an absolute improvement in few-shot accuracy of up to 30% in 9 math tasks. After fine-tuning, R HO -1-1B and 7B achieved state-of-the-art results of 40.6% and 51.8% on MATH dataset, respectively — matching DeepSeekMath with only 3% of the pretraining tokens. Furthermore, when continual pretraining on 80B general tokens, R HO -1 achieves 6.8% average enhancement across 15 diverse tasks, increasing both data efficiency and performance of the language model pre-training.