tkitsers/Llemma-metamath-7b

This model is Llemma-7b model used in the paper "An Empirical Analysis of Compute-Optimal Inference for Problem-Solving with Language Models". It's based on Llemma-7b and was further finetuned MetaMath with special format for reward. Each step starts with "Step" and ends with "\u043a\u0438".