Nous Hermes 2 Mistral 7B

About

The Nous Hermes 2 Mistral 7B is a large language model developed by NousResearch, utilizing the Mistral 7B architecture. It features an extensive context window, making it adept at general-purpose language processing, role-playing, and creative writing. Excelling in rapid data processing and basic tasks, it incorporates Direct Preference Optimization (DPO) for adaptive learning through human feedback. The model supports function calling and JSON mode with high accuracy, although its complex dialogue handling and domain knowledge have limitations. Licensed under Apache 2.0, it showcases improved performance on benchmarks like AGIEval and TruthfulQA.

Capabilities

Providers(1)