allenai/llama-3-tulu-2-8b-uf-mean-rm

<center> <img src="https://huggingface.co/datasets/allenai/blog-images/resolve/main/tulu-2.5/tulu25banner.png" alt="Tulu 2.5 banner image" width="800px"/> </center>

Model Card for Llama 3 Tulu V2 8B RM - UltraFeedback

Tulu is a series of language models that are trained to act as helpful assistants. This is a 8B reward model used for PPO training trained on the UltraFeedback dataset.

For more details, read the paper: Unpacking DPO and PPO: Disentangling Best Practices for Learning from Preference Feedback.

Built with Meta Llama 3! Note that Llama 3 is released under the Meta Llama 3 community license, included here under llama_3_license.txt.

Performance

We evaluate the model on RewardBench:

Model description

• —Model type: A reward model trained on UltraFeedback, designed to be used in RLHF training.
• —Language(s) (NLP): English
• —License: Apache 2.0.
• —Finetuned from model: allenai/llama-3-tulu-2-8b

Model Sources

• —Repository: https://github.com/allenai/open-instruct
• —Dataset: Data used to train this model can be found here - specifically the ultrafeedback_mean_aspects split.

Input Format

The model is trained to use the following format (note the newlines):

<|user|>
Your message here!
<|assistant|>

For best results, format all inputs in this manner. Make sure to include a newline after `<|assistant|>`, this can affect generation quality quite a bit. We have included a chat template in the tokenizer implementing this template.

Intended uses & limitations

The model was initially fine-tuned on a filtered and preprocessed of the Tulu V2 mix dataset, which contains a diverse range of human created instructions and synthetic dialogues generated primarily by other LLMs. We then further trained the model with a Jax RM trainer built on EasyLM on the dataset mentioned above. This model is meant as a research artefact.

Training hyperparameters

The following hyperparameters were used during PPO training:

• —learning_rate: 1e-06
• —totaltrainbatch_size: 512
• —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• —lrschedulertype: linear cooldown to 1e-05.
• —lrschedulerwarmup_ratio: 0.03
• —num_epochs: 1.0

Citation

If you find Tulu 2.5 is useful in your work, please cite it with:

@misc{ivison2024unpacking,
      title={{Unpacking DPO and PPO: Disentangling Best Practices for Learning from Preference Feedback}}, 
      author={{Hamish Ivison and Yizhong Wang and Jiacheng Liu and Ellen Wu and Valentina Pyatkin and Nathan Lambert and Yejin Choi and Noah A. Smith and Hannaneh Hajishirzi}}
      year={2024},
      eprint={2406.09279},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}