metadata

base_model:
  - chargoddard/prometheus-llama-3-8b-preference
  - chargoddard/prometheus-llama-3-8b-absolute
library_name: transformers
tags:
  - mergekit
  - merge
license: apache-2.0
datasets:
  - prometheus-eval/Preference-Collection
  - prometheus-eval/Feedback-Collection
language:
  - en

prometheus-2-llama-3-8b

Replication of prometheus-7b-v2.0 using Llama 3 8B Instruct as a base model.

As in their paper, two different models were trained on their preference and feedback datasets then linearly merged at equal weight.

Training hyperparameters:

1 epoch
Learning rate 1e-5
Effective batch size 128
Cosine annealing
~5% warmup

Uses Llama 3 Instruct prompt format and the same prompts as prometheus-7b-v2.0. See that readme for info.

Citations

@misc{kim2023prometheus,
    title={Prometheus: Inducing Fine-grained Evaluation Capability in Language Models},
    author={Seungone Kim and Jamin Shin and Yejin Cho and Joel Jang and Shayne Longpre and Hwaran Lee and Sangdoo Yun and Seongjin Shin and Sungdong Kim and James Thorne and Minjoon Seo},
    year={2023},
    eprint={2310.08491},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}

@misc{kim2024prometheus,
    title={Prometheus 2: An Open Source Language Model Specialized in Evaluating Other Language Models},
    author={Seungone Kim and Juyoung Suk and Shayne Longpre and Bill Yuchen Lin and Jamin Shin and Sean Welleck and Graham Neubig and Moontae Lee and Kyungjae Lee and Minjoon Seo},
    year={2024},
    eprint={2405.01535},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}