Summary of Infobench: Evaluating Instruction Following Ability in Large Language Models, by Yiwei Qin et al.
InFoBench: Evaluating Instruction Following Ability in Large Language Models
by Yiwei Qin, Kaiqiang Song, Yebowen Hu, Wenlin Yao, Sangwoo Cho, Xiaoyang Wang, Xuansheng Wu, Fei Liu, Pengfei Liu, Dong Yu
First submitted to arxiv on: 7 Jan 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 This paper introduces the Decomposed Requirements Following Ratio (DRFR), a new metric that evaluates Large Language Models’ (LLMs) ability to follow instructions by breaking down complex tasks into simpler criteria. The authors present InFoBench, a benchmark comprising 500 diverse instructions and 2,250 decomposed questions across multiple constraint categories. The study compares DRFR with traditional scoring methods and explores annotation sources, including human experts, crowd-sourced workers, and GPT-4. The results show that DRFR is more reliable than traditional methods and that GPT-4 can be used as a cost-efficient annotator. The evaluation of several advanced LLMs using this framework reveals their strengths and areas needing improvement. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper creates a new way to measure how well Large Language Models (LLMs) follow instructions. It’s like giving them a list of tasks and seeing if they can do each one correctly. The researchers made a special tool called the Decomposed Requirements Following Ratio (DRFR) that helps break down big tasks into smaller ones, making it easier to understand what the LLMs are doing right or wrong. They also created a big dataset with lots of different instructions and questions to test the LLMs on. This study is important because it helps us make better language models in the future. |
Keywords
» Artificial intelligence » Gpt
