[GRPO Explained] DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models

#deepseek #llm #grpo GRPO is one of the core advancements used in Deepseek-R1, but was introduced already last year in this paper that uses a combination of new RL techniques and iterative data collection to achieve remarkable performance on mathematics benchmarks with just a 7B model. Paper: https://arxiv.org/abs/2402.03300 Abstract: Mathematical reasoning poses a significant challenge for language models due to its complex and structured nature. In this paper, we introduce DeepSeekMath 7B, which continues pre-training DeepSeek-Coder-Base-v1.5 7B with 120B math-related tokens sourced from Common Crawl, together with natural language and code data. DeepSeekMath 7B has achieved an impressive score of 51.7% on the competition-level MATH benchmark without relying on external toolkits and voting techniques, approaching the performance level of Gemini-Ultra and GPT-4. Self-consistency over 64 samples from DeepSeekMath 7B achieves 60.9% on MATH. The mathematical reasoning capability of DeepS