@inproceedings{yoon-etal-2023-gradient,
    title = "Gradient Ascent Post-training Enhances Language Model Generalization",
    author = "Yoon, Dongkeun  and
      Jang, Joel  and
      Kim, Sungdong  and
      Seo, Minjoon",
    editor = "Rogers, Anna  and
      Boyd-Graber, Jordan  and
      Okazaki, Naoaki",
    booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)",
    month = jul,
    year = "2023",
    address = "Toronto, Canada",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-emnlp/2023.acl-short.74/",
    doi = "10.18653/v1/2023.acl-short.74",
    pages = "851--864",
    abstract = "In this work, we empirically show that updating pretrained LMs (350M, 1.3B, 2.7B) with just a few steps of Gradient Ascent Post-training (GAP) on random, unlabeled text corpora enhances its zero-shot generalization capabilities across diverse NLP tasks. Specifically, we show that GAP can allow LMs to become comparable to 2-3x times larger LMs across 12 different NLP tasks. We also show that applying GAP on out-of-distribution corpora leads to the most reliable performance improvements. Our findings indicate that GAP can be a promising method for improving the generalization capability of LMs without any task-specific fine-tuning."
}