Summary of Quill: Quotation Generation Enhancement Of Large Language Models, by Jin Xiao et al.
QUILL: Quotation Generation Enhancement of Large Language Models
by Jin Xiao, Bowei Zhang, Qianyu He, Jiaqing Liang, Feng Wei, Jinglei Chen, Zujie Liang, Deqing Yang, Yanghua Xiao
First submitted to arxiv on: 6 Nov 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
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 The abstract proposes a comprehensive approach to evaluate and improve the performance of Large Language Models (LLMs) in generating quotations. The authors develop a holistic evaluation system for quotation generation tasks, consisting of five criteria with corresponding automatic metrics. To enhance LLMs’ abilities, they construct a bilingual knowledge base containing 32,022 quotes and design a quotation-specific metric to rerank retrieved quotes. Experimental results demonstrate strong correlation between proposed metrics and human preferences, indicating that existing LLMs struggle to generate desired quotes, but the proposed approach helps bridge this gap. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Large Language Models are great at helping us write, but they’re not very good at generating quotes that sound like real people. They either make things up or can’t do it as well as humans. To fix this, researchers developed a way to measure how well LLMs do at generating quotes and then made some changes to help them do better. They created a huge database of quotes in two languages and came up with new ways to rank the quotes that LLMs find. The results show that the new approach works well and helps LLMs generate more realistic quotes. |
Keywords
» Artificial intelligence » Knowledge base
