EleutherAI/llemma_7b

EleutherAI's Llemma 7B is a 7 billion parameter language model specifically designed for mathematics. Initialized from Code Llama 7B, it was further trained on the Proof-Pile-2 dataset for 200 billion tokens. This model excels at chain-of-thought mathematical reasoning and is adept at using computational tools for mathematical tasks, outperforming Llama-2 and Code Llama on various math benchmarks.