NAACL 2024short0 citations

On Retrieval Augmentation and the Limitations of Language Model Training

Ting-Rui Chiang, Xinyan Yu, Joshua Robinson, Ollie Liu, Isabelle Lee, Dani Yogatama

Abstract

Augmenting a language model (LM) with k-nearest neighbors (kNN) retrieval on its training data alone can decrease its perplexity, though the underlying reasons for this remain elusive. In this work, we rule out one previously posited possibility — the “softmax bottleneck.” We then create a new dataset to evaluate LM generalization ability in the setting where training data contains additional information that is not causally relevant. This task is challenging even for GPT-3.5 Turbo. We show that, for both GPT-2 and Mistral 7B, kNN retrieval augmentation consistently improves per formance in this setting. Finally, to make kNN retrieval more accessible, we propose using amulti-layer perceptron model that maps datastore keys to values as a drop-in replacement for traditional retrieval. This reduces storage costsby over 25x.

BibTeX
@inproceedings{chiang-etal-2024-retrieval,
    title = "On Retrieval Augmentation and the Limitations of Language Model Training",
    author = "Chiang, Ting-Rui  and
      Yu, Xinyan  and
      Robinson, Joshua  and
      Liu, Ollie  and
      Lee, Isabelle  and
      Yogatama, Dani",
    editor = "Duh, Kevin  and
      Gomez, Helena  and
      Bethard, Steven",
    booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers)",
    month = jun,
    year = "2024",
    address = "Mexico City, Mexico",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2024.naacl-short.21/",
    doi = "10.18653/v1/2024.naacl-short.21",
    pages = "229--238"
}
On Retrieval Augmentation and the Limitations of Language Model Training · NAACL 2024