References & Citations
Computer Science > Computation and Language
Title: LLMRefine: Pinpointing and Refining Large Language Models via Fine-Grained Actionable Feedback
(Submitted on 15 Nov 2023 (v1), last revised 2 Apr 2024 (this version, v3))
Abstract: Recent large language models (LLM) are leveraging human feedback to improve their generation quality. However, human feedback is costly to obtain, especially during inference. In this work, we propose LLMRefine, an inference time optimization method to refine LLM's output. The core idea is to use a learned fine-grained feedback model to pinpoint defects and guide LLM to refine them iteratively. Using original LLM as a proposal of edits, LLMRefine searches for defect-less text via simulated annealing, trading off the exploration and exploitation. We conduct experiments on three text generation tasks, including machine translation, long-form question answering (QA), and topical summarization. LLMRefine consistently outperforms all baseline approaches, achieving improvements up to 1.7 MetricX points on translation tasks, 8.1 ROUGE-L on ASQA, 2.2 ROUGE-L on topical summarization.
Submission history
From: Wenda Xu [view email][v1] Wed, 15 Nov 2023 19:52:11 GMT (1508kb,D)
[v2] Thu, 28 Mar 2024 00:50:55 GMT (1802kb,D)
[v3] Tue, 2 Apr 2024 16:39:13 GMT (2540kb,D)
Link back to: arXiv, form interface, contact.