Summary of Fanoutqa: a Multi-hop, Multi-document Question Answering Benchmark For Large Language Models, by Andrew Zhu and Alyssa Hwang and Liam Dugan and Chris Callison-burch
FanOutQA: A Multi-Hop, Multi-Document Question Answering Benchmark for Large Language Models
by Andrew Zhu, Alyssa Hwang, Liam Dugan, Chris Callison-Burch
First submitted to arxiv on: 21 Feb 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 research paper presents FanOutQA, a high-quality dataset for evaluating the complex reasoning capabilities of large language models (LLMs) in answering “fan-out” questions. These multi-hop, multi-document reasoning questions require finding information about multiple entities across a large number of documents. The authors benchmark 7 LLMs on three settings across their dataset, which includes GPT-4, LLaMA 2, Claude-2.1, and Mixtral-8x7B. Results show that contemporary models still have room for improvement in reasoning over inter-document dependencies. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary FanOutQA is a new way to test how well language models can understand and answer complex questions. These questions are like puzzles that need to be solved by finding information from many different sources. The researchers created a big dataset of these types of questions, along with the answers, and then tested several language models on this dataset. They found that even the best language models still have trouble figuring out how to use information from multiple documents. |
Keywords
» Artificial intelligence » Claude » Gpt » Llama
