Saved in:
Bibliographic Details
Main Authors: Antoine, Elie, Béchet, Frédéric, Damnati, Géraldine, Langlais, Philippe
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2501.17569
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916588859424768
author Antoine, Elie
Béchet, Frédéric
Damnati, Géraldine
Langlais, Philippe
author_facet Antoine, Elie
Béchet, Frédéric
Damnati, Géraldine
Langlais, Philippe
contents We introduce an evaluation methodology for reading comprehension tasks based on the intuition that certain examples, by the virtue of their linguistic complexity, consistently yield lower scores regardless of model size or architecture. We capitalize on semantic frame annotation for characterizing this complexity, and study seven complexity factors that may account for model's difficulty. We first deploy this methodology on a carefully annotated French reading comprehension benchmark showing that two of those complexity factors are indeed good predictors of models' failure, while others are less so. We further deploy our methodology on a well studied English benchmark by using Chat-GPT as a proxy for semantic annotation. Our study reveals that fine-grained linguisticallymotivated automatic evaluation of a reading comprehension task is not only possible, but helps understand models' abilities to handle specific linguistic characteristics of input examples. It also shows that current state-of-the-art models fail with some for those characteristics which suggests that adequately handling them requires more than merely increasing model size.
format Preprint
id arxiv_https___arxiv_org_abs_2501_17569
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle A linguistically-motivated evaluation methodology for unraveling model's abilities in reading comprehension tasks
Antoine, Elie
Béchet, Frédéric
Damnati, Géraldine
Langlais, Philippe
Computation and Language
We introduce an evaluation methodology for reading comprehension tasks based on the intuition that certain examples, by the virtue of their linguistic complexity, consistently yield lower scores regardless of model size or architecture. We capitalize on semantic frame annotation for characterizing this complexity, and study seven complexity factors that may account for model's difficulty. We first deploy this methodology on a carefully annotated French reading comprehension benchmark showing that two of those complexity factors are indeed good predictors of models' failure, while others are less so. We further deploy our methodology on a well studied English benchmark by using Chat-GPT as a proxy for semantic annotation. Our study reveals that fine-grained linguisticallymotivated automatic evaluation of a reading comprehension task is not only possible, but helps understand models' abilities to handle specific linguistic characteristics of input examples. It also shows that current state-of-the-art models fail with some for those characteristics which suggests that adequately handling them requires more than merely increasing model size.
title A linguistically-motivated evaluation methodology for unraveling model's abilities in reading comprehension tasks
topic Computation and Language
url https://arxiv.org/abs/2501.17569