mzhaoshuai/Mistral-7B-v0.1-conf-refalign

RefAlign: RL with Similarity-based Rewards

GitHub repository: https://github.com/mzhaoshuai/RefAlign

Paper: Learning from Reference Answers: Versatile Language Model Alignment without Binary Human Preference Data.

This repository hosts mzhaoshuai/Mistral-7B-v0.1-conf-refalign, a model fine-tuned for confidence alignment using the RefAlign framework. It is based on mzhaoshuai/Mistral-7B-v0.1-conf-sft and presented in the paper Learning from Reference Answers: Versatile Language Model Alignment without Binary Human Preference Data.

Abstract

Large language models~(LLMs) are expected to be helpful, harmless, and honest. In different alignment scenarios, such as safety, confidence, and general preference alignment, binary preference data collection and reward modeling are resource-intensive but play a central role in transferring human preferences. In this work, we explore using the similarity between sampled generations and reference answers as a supplementary reward function for alignment. When unary reference answers are available, such similarity-based rewards can circumvent the need for binary preference data and explicit reward modeling. We introduce \textit{RefAlign}, a versatile REINFORCE-style alignment algorithm that does not rely on reward or reference models. RefAlign utilizes language generation evaluation metrics, such as BERTScore, between sampled generations and reference answers as surrogate rewards. Beyond general preference optimization, RefAlign can be naturally extended to diverse scenarios, including safety and confidence alignment, by combining similarity-based rewards with task-specific objectives. Across multiple scenarios, RefAlign achieves performance comparable to prior alignment methods while operating without binary preference data or reward models.

Framework versions

• PEFT 0.11.1

Bibtex

@article{zhao2025learning,
  title={Learning from reference answers: Versatile language model alignment without binary human preference data},
  author={Zhao, Shuai and Xu, Yunqiu and Zhu, Linchao and Yang, Yi},
  journal={arXiv preprint arXiv:2504.09895},
  year={2025}
}