prometheus-eval/prometheus-7b-v2.0

Prometheus 2 (prometheus-eval/prometheus-7b-v2.0) is a 7 billion parameter language model developed by prometheus-eval, built upon the Mistral-Instruct base architecture with a 4096 token context length. It is specifically fine-tuned on 100K feedback and 200K preference data for fine-grained evaluation of other LLMs and functions as a reward model for RLHF. This model uniquely supports both absolute grading (direct assessment) and relative grading (pairwise ranking) through weight merging, making it an alternative to GPT-4 for detailed LLM evaluation.