LongReward: Improving Long-context Large Language Models with AI Feedback
This addresses the challenge of obtaining reliable rewards for long-context scenarios in LLMs, offering a novel approach to enhance model performance, though it is incremental as it builds on existing RL and SFT methods.
The paper tackles the problem of compromised data quality in long-context large language models (LLMs) by proposing LongReward, a method that uses an off-the-shelf LLM to provide rewards from four human-valued dimensions, combined with DPO to improve long-context SFT models, resulting in significant performance improvements in long-context tasks and enhanced short-instruction following.
Though significant advancements have been achieved in developing long-context large language models (LLMs), the compromised quality of LLM-synthesized data for supervised fine-tuning (SFT) often affects the long-context performance of SFT models and leads to inherent limitations. In principle, reinforcement learning (RL) with appropriate reward signals can further enhance models' capacities. However, how to obtain reliable rewards in long-context scenarios remains unexplored. To this end, we propose LongReward, a novel method that utilizes an off-the-shelf LLM to provide rewards for long-context model responses from four human-valued dimensions: helpfulness, logicality, faithfulness, and completeness, each with a carefully designed assessment pipeline. By combining LongReward and offline RL algorithm DPO, we are able to effectively improve long-context SFT models. Our experiments indicate that LongReward not only significantly improves models' long-context performance but also enhances their ability to follow short instructions. We also find that long-context DPO with LongReward and conventional short-context DPO can be used together without hurting either one's performance.