Training Language Models to Generate Text with Citations via Fine-grained Rewards
This addresses the issue of unreliable LLM responses for users needing credible information, but it is incremental as it builds on existing citation methods.
The authors tackled the problem of LLMs hallucinating and lacking credibility by training them to generate text with in-text citations using fine-grained rewards, achieving the best performance on LLaMA-2-7B and surpassing GPT-3.5-turbo on QA datasets.
While recent Large Language Models (LLMs) have proven useful in answering user queries, they are prone to hallucination, and their responses often lack credibility due to missing references to reliable sources. An intuitive solution to these issues would be to include in-text citations referring to external documents as evidence. While previous works have directly prompted LLMs to generate in-text citations, their performances are far from satisfactory, especially when it comes to smaller LLMs. In this work, we propose an effective training framework using fine-grained rewards to teach LLMs to generate highly supportive and relevant citations, while ensuring the correctness of their responses. We also conduct a systematic analysis of applying these fine-grained rewards to common LLM training strategies, demonstrating its advantage over conventional practices. We conduct extensive experiments on Question Answering (QA) datasets taken from the ALCE benchmark and validate the model's generalizability using EXPERTQA. On LLaMA-2-7B, the incorporation of fine-grained rewards achieves the best performance among the baselines, even surpassing that of GPT-3.5-turbo.