학술논문
CRASS: A Novel Data Set and Benchmark to Test Counterfactual Reasoning of Large Language Models
Document Type
Working Paper
Author
Source
Proceedings of the 13th Language Resources and Evaluation Conference (LREC 2022), Marseille, France pp. 2126-2140 (2022) https://aclanthology.org/2022.lrec-1.229/
Subject
Language
Abstract
We introduce the CRASS (counterfactual reasoning assessment) data set and benchmark utilizing questionized counterfactual conditionals as a novel and powerful tool to evaluate large language models. We present the data set design and benchmark that supports scoring against a crowd-validated human baseline. We test six state-of-the-art models against our benchmark. Our results show that it poses a valid challenge for these models and opens up considerable room for their improvement.
Comment: 10 pages including references, plus 5 pages appendix. Edits for version 3 vs LREC 2022: Point out human baseline in abstract (also to match arxiv abstract), fix affiliation apergo.ai, and fix a recurring typo
Comment: 10 pages including references, plus 5 pages appendix. Edits for version 3 vs LREC 2022: Point out human baseline in abstract (also to match arxiv abstract), fix affiliation apergo.ai, and fix a recurring typo