LLM2Vec Meta Llama 3 8B Instruct MNTP

An 8-billion parameter embedding model from McGill NLP, converting Meta's Llama 3 Instruct into a text embedding model via masked next token prediction.