@inproceedings{stefanik-etal-2025-distribution,
    title = "Can Out-of-Distribution Evaluations Uncover Reliance on Prediction Shortcuts? A Case Study in Question Answering",
    author = "{\v{S}}tef{\'a}nik, Michal  and
      Mickus, Timothee  and
      Spiegel, Michal  and
      Kadl{\v{c}}{\'i}k, Marek  and
      Kucha{\v{r}}, Josef",
    editor = "Christodoulopoulos, Christos  and
      Chakraborty, Tanmoy  and
      Rose, Carolyn  and
      Peng, Violet",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2025",
    month = nov,
    year = "2025",
    address = "Suzhou, China",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-luhme/2025.findings-emnlp.1232/",
    doi = "10.18653/v1/2025.findings-emnlp.1232",
    pages = "22628--22635",
    ISBN = "979-8-89176-335-7",
    abstract = "A large body of recent work assesses models' generalization capabilities through the lens of performance on out-of-distribution (OOD) datasets. Despite their practicality, such evaluations build upon a strong assumption: that OOD evaluations can capture and reflect upon possible failures in a real-world deployment. In this work, we challenge this assumption and confront the results obtained from OOD evaluations with a set of specific failure modes documented in existing question-answering (QA) models, referred to as a reliance on spurious features or prediction shortcuts.We find that different datasets used for OOD evaluations in QA provide an estimate of models' robustness to shortcuts that have a vastly different quality, some largely under-performing even a simple, in-distribution evaluation. We partially attribute this to the observation that spurious shortcuts are shared across ID+OOD datasets, but also find cases where a dataset{'}s quality for training and evaluation is largely disconnected. Our work underlines limitations of commonly-used OOD-based evaluations of generalization, and provides methodology and recommendations for evaluating generalization within and beyond QA more robustly."
}