microsoft/rho-math-7b-v0.1
TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:4kPublished:Apr 11, 2024License:mitArchitecture:Transformer0.0K Open Weights Cold

The microsoft/rho-math-7b-v0.1 model is a 7 billion parameter language model developed by Microsoft, pre-trained using Selective Language Modeling (SLM) to optimize for mathematical reasoning. This approach allows it to achieve strong performance on math benchmarks like MATH and GSM8K with significantly fewer pretraining tokens compared to traditional methods. It demonstrates 31.0% few-shot accuracy on the MATH dataset and 66.9% on GSM8K, making it suitable for applications requiring efficient and accurate mathematical problem-solving.

Loading preview...