Summary of Benchmarking Complex Instruction-following with Multiple Constraints Composition, by Bosi Wen et al.
Benchmarking Complex Instruction-Following with Multiple Constraints Composition
by Bosi Wen, Pei Ke, Xiaotao Gu, Lindong Wu, Hao Huang, Jinfeng Zhou, Wenchuang Li, Binxin Hu, Wendy Gao, Jiaxin Xu, Yiming Liu, Jie Tang, Hongning Wang, Minlie Huang
First submitted to arxiv on: 4 Jul 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 proposed paper introduces a new benchmark called ComplexBench for evaluating the ability of large language models (LLMs) to follow complex human instructions composed of multiple constraints. The authors argue that current benchmarks focus primarily on modeling different types of constraints, neglecting the composition of these constraints in real-world scenarios. To address this gap, they propose a hierarchical taxonomy for complex instructions, including constraint types, dimensions, and composition types, and manually collect a high-quality dataset accordingly. The evaluation process is based on LLM-based evaluators that verify whether generated texts satisfy each constraint and composition, with the final score determined by dependency structure. This paper highlights the significant deficiencies in existing LLMs when dealing with complex instructions with multiple constraints composition. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Large language models can do many things, but one important thing they’re getting better at is following instructions. Instructions are like recipes for making something, and language models need to be able to understand what’s being asked of them. But right now, there aren’t good ways to test how well language models can follow complex instructions. That’s why the authors of this paper created a new benchmark called ComplexBench. This benchmark helps evaluate how well language models can follow complex instructions with many rules and conditions. The authors also came up with a way to classify these complex instructions into different categories, so it’s easier to test language models on them. |
