Examining the robustness of LLM evaluation to the distributional assumptions of benchmarks

Charlotte Siska, Katerina Marazopoulou, Melissa Ailem, James Bono


Abstract
Benchmarks have emerged as the central approach for evaluating Large Language Models (LLMs). The research community often relies on a model’s average performance across the test prompts of a benchmark to evaluate the model’s performance. This is consistent with the assumption that the test prompts within a benchmark represent a random sample from some real-world distribution of interest. We note that this is generally not the case; instead, we hold that the distribution of interest varies according to the specific use case. Hence, we analyze the robustness of LLM benchmarks to their underlying distributional assumptions. We find that (1) the correlation in model performance across test prompts is non-random, (2) accounting for correlations across test prompts can change model rankings on major benchmarks, (3) explanatory factors for these correlations include semantic similarity and common LLM failure points.
Anthology ID:
2024.acl-long.560
Volume:
Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
Month:
August
Year:
2024
Address:
Bangkok, Thailand
Editors:
Lun-Wei Ku, Andre Martins, Vivek Srikumar
Venue:
ACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
10406–10421
Language:
URL:
https://aclanthology.org/2024.acl-long.560
DOI:
10.18653/v1/2024.acl-long.560
Bibkey:
Cite (ACL):
Charlotte Siska, Katerina Marazopoulou, Melissa Ailem, and James Bono. 2024. Examining the robustness of LLM evaluation to the distributional assumptions of benchmarks. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 10406–10421, Bangkok, Thailand. Association for Computational Linguistics.
Cite (Informal):
Examining the robustness of LLM evaluation to the distributional assumptions of benchmarks (Siska et al., ACL 2024)
Copy Citation:
PDF:
https://preview.aclanthology.org/autopr/2024.acl-long.560.pdf