Smaug: Fixing Failure Modes of Preference Optimisation with DPO-Positive

TL;DR

This paper introduces DPO-Positive, a new loss function that addresses failure modes in preference-based fine-tuning of large language models, leading to improved performance across various tasks and datasets.

Contribution

The authors identify a failure mode in standard DPO and propose DPO-Positive, a novel training method that outperforms DPO and other fine-tuning techniques on multiple benchmarks.

Findings

DPO can reduce the likelihood of preferred examples during training.

DPO-Positive mitigates this issue and improves downstream task performance.

Smaug-72B surpasses 80% accuracy on HuggingFace leaderboard.

Abstract

Direct Preference Optimisation (DPO) is effective at significantly improving the performance of large language models (LLMs) on downstream tasks such as reasoning, summarisation, and alignment. Using pairs of preferred and dispreferred data, DPO models the relative probability of picking one response over another. In this work, first we show theoretically that the standard DPO loss can lead to a reduction of the model's likelihood of the preferred examples, as long as the relative probability between the preferred and dispreferred classes increases. We then show empirically that this phenomenon occurs when fine-tuning LLMs on common datasets, especially datasets in which the edit distance between pairs of completions is low. Using these insights, we design DPO-Positive (DPOP), a new loss function and training procedure which avoids this failure mode. Surprisingly, we find that DPOP outperforms DPO and other fine-tuning procedures across a wide variety of datasets and downstream tasks, including datasets with high edit distances between completions. Furthermore, we find that the DPOP-tuned model outperforms the DPO-tuned model (all else equal) on benchmarks independent of the fine-tuning data, such as MT-Bench. Finally, using DPOP, we create and open-source Smaug-34B and Smaug-72B, with the latter becoming the first open-source LLM to surpass an average accuracy of 80% on the HuggingFace Open LLM Leaderboard.