Reinforcement Learning without Human Feedback for Last Mile Fine-Tuning of Large Language Models

TL;DR

This paper introduces a reinforcement learning framework for last-mile fine-tuning of large language models, improving performance over traditional likelihood maximization, especially in tasks like abstractive summarization.

Contribution

It develops a novel reinforcement learning-based approach for last-mile fine-tuning, demonstrating its effectiveness beyond likelihood maximization in language model optimization.

Findings

Reinforcement learning outperforms likelihood maximization in raw prediction quality.

The performance gap can be reduced with post-processing of likelihood outputs.

Framework is adaptable to penalize undesirable outputs like hallucinations.

Abstract

Reinforcement learning is used to align language models with human preference signals after first pre-training the model to predict the next token of text within a large corpus using likelihood maximization. Before being deployed in a specific domain, models are often further fine-tuned on task specific data. Since human preferences are often unavailable for the last step, it is performed using likelihood maximization as that is the typical default method. However, reinforcement learning has other advantages besides facilitating alignment to a human derived reward function. For one, whereas likelihood maximization is a form of imitation learning in which the model is trained on what to do under ideal conditions, reinforcement learning is not limited to demonstrating actions just for optimally reached states and trains a model what to do under a range of scenarios as it explores the policy space. In addition, it also trains a model what not to do, suppressing competitive but poor actions. This work develops a framework for last-mile fine-tuning using reinforcement learning and tests whether it garners performance gains. The experiments center on abstractive summarization, but the framework is general and broadly applicable. Use of the procedure produced significantly better results than likelihood maximization when comparing raw predictions. For the specific data tested, the gap could be bridged by employing post-processing of the maximum likelihood outputs. Nonetheless, the framework offers a new avenue for model optimization in situations where post-processing may be less straightforward or effective, and it can be extended to include more complex classes of undesirable outputs to penalize and train against, such as hallucinations.