MetaMath Llemma 7B
MetaMath Llemma 7B has model metadata, but missing tracked provider pricing keeps it from being a default production pick.
Use it for
- Teams evaluating general LLM work
Do not use it for
- Cost-sensitive launches that need sourced token pricing
- Vision or document-understanding workloads
- Strict JSON or tool-calling flows
- Family
- MetaMath
- Released
- 2023-10-27
- Parameters
- 7B
- Architecture
- Decoder Only
- Specialization
- general
- Training
- finetuned
About
MetaMath-Llemma-7B is a fine-tuned large language model designed specifically for mathematical reasoning, building upon the Llemma-7B architecture. It showcases notably enhanced performance over similar-sized open-source models in solving mathematical problems, thanks to a novel question bootstrapping method that enriches its training data with diverse mathematical questions. Trained on the comprehensive MetaMathQA dataset, which is derived from GSM8K and MATH benchmarks, the model achieves impressive scores: Pass@1 scores of 69.2% on GSM8K and 30.0% on MATH. Despite its achievements, it remains outperformed by closed-source models like GPT-4, pointing to avenues for further improvement. Additionally, its reliance on datasets augmented with tools like ChatGPT 3.5 may introduce certain biases. Nevertheless, its open-source nature allows for ongoing research and improvements by the AI community.
MetaMath Llemma 7B is a model in the MetaMath family. No headline benchmark score is tracked for MetaMath Llemma 7B yet.
Top use-case fit
No primary decision-task fit is mapped for this model yet.
Provider price ladder
No tracked provider token pricing is available for this model yet.
Capabilities
No model capability flags are currently sourced.
Benchmark peer barsfor Coding
No task-mapped benchmark peers are available for this model yet.
Migration checks
No linked migration route is available for this model yet.