Instructed Models
Mixtral released Mixtral 8x7B Instruct alongside Mixtral 8x7B. This model has been optimized through supervised fine-tuning and direct preference optimization (DPO) for careful instruction following. On MT-Bench, it reaches a score of 8.30, making it the best open-source model, with a performance comparable to GPT-3.5.
Note: Mixtral can be gracefully prompted to ban some outputs for constructing applications that require a strong level of moderation, as exemplified here. Proper preference tuning can also serve this purpose. Bear in mind that without such a prompt, the model will follow whatever instructions are given.
Deploy Mixtral with an Open-Source Deployment Stack
To enable the community to run Mixtral with a fully open-source stack, they have submitted changes to the vLLM project, which integrates Megablocks CUDA kernels for efficient inference.