EleutherAI/math-lm

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1097 stars · 95 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Repository for Llemma, an open language model for mathematics, hosting training code, data preprocessing scripts, and evaluation tooling. It includes the Proof-Pile-2 and AlgebraicStack datasets and links to the Llemma 7B and 34B model weights on Hugging Face.

Use cases

• train a language model for math reasoning
• preprocess mathematical web text for LLM training
• evaluate LLMs on math benchmarks
• fine-tune a model on mathematical text
• run theorem proving experiments
• download the proof-pile-2 dataset

When to choose

• you need an open math-focused LLM or its training data
• you want to reproduce or extend the Llemma training pipeline
• you need curated mathematical pretraining datasets

When to avoid

• you just want to run inference with a math LLM without training
• you need a production-ready, well-packaged library
• you lack GPU training infrastructure

Facets

library · maturity maintenance

llm-training machine-learning data-generation benchmarking large-language-models mathematics machine-learning python llemma math-reasoning gpt-neox proof-pile-2 research-code eleutherai natural-language-processing gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="EleutherAI/math-lm")