# Safety alignment

Minerva is an Italian-English bilingual large language model developed by Sapienza NLP, based on the Mistral architecture with 7 billion parameters, supporting instruction-following and dialogue tasks.

Aloe is a brand-new family of large medical language models. Through model fusion and advanced prompting strategies, it is highly competitive among open-source models of the same type.

A reward model trained on Meta-Llama-3-8B-Instruct for reward modeling in RLHF processes, supporting PPO, iterative SFT, and iterative DPO methods.

Vistral-7B-Chat is a large language model for Vietnamese multi-turn conversations, developed based on the Mistral 7B architecture with special optimizations for Vietnamese language processing