Summary of On the Capacity Of Citation Generation by Large Language Models, By Haosheng Qian et al.
On the Capacity of Citation Generation by Large Language Models
by Haosheng Qian, Yixing Fan, Ruqing Zhang, Jiafeng Guo
First submitted to arxiv on: 15 Oct 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper presents a novel approach to alleviating the “hallucination” problem in large language models (LLMs) by incorporating external traceable resources for response generation. The core idea is to accurately attribute claims in responses to the corresponding retrieved documents, which existing works have largely overlooked. The study systematically analyzes the capabilities of LLMs in generating citations within response generation and introduces a novel method to enhance their citation generation abilities. The evaluation focuses on two benchmark datasets, using new citation evaluation metrics that eliminate over-penalization of unnecessary and excessive citations. Additionally, the paper proposes a Generate-then-Refine method that completes relevant citations and removes irrelevant ones without altering the response text. Experimental results on WebGLM-QA, ASQA, and ELI5 datasets demonstrate substantial improvements in citation quality. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study helps us understand how to improve the way large language models create responses by giving credit to the sources they use. The researchers looked at seven popular language models and tested them on two big datasets. They came up with a new way for these models to generate citations that are accurate and helpful, rather than just listing lots of unnecessary references. This could make it easier for people to find reliable information online and trust what they read. |
Keywords
» Artificial intelligence » Hallucination
