Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
80 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
7 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

More Room for Language: Investigating the Effect of Retrieval on Language Models (2404.10939v1)

Published 16 Apr 2024 in cs.CL

Abstract: Retrieval-augmented LLMs pose a promising alternative to standard LLMing. During pretraining, these models search in a corpus of documents for contextually relevant information that could aid the LLMing objective. We introduce an 'ideal retrieval' methodology to study these models in a fully controllable setting. We conduct an extensive evaluation to examine how retrieval augmentation affects the behavior of the underlying LLM. Among other things, we observe that these models: i) save substantially less world knowledge in their weights, ii) are better at understanding local context and inter-word dependencies, but iii) are worse at comprehending global context.

PDF HTML Abstract

Exploration of Retrieval-Augmented Pretraining for LLMs

Introduction

Retrieval-augmented LLMs leverage both self-supervised learning and external information retrieval to enhance their ability to generate contextually relevant responses. These models integrate a nonparametric memory in the form of data retrieval during the token prediction process in training, which theoretically aids the model by providing additional context from a knowledge database. Several studies have shown the efficacy of these models in specific tasks like open-domain question answering. However, their impact on the core functionalities and behaviors of the underlying LLMs, when isolated from the retrieval components, is less studied.

Methodology

The paper introduces a structured methodology to evaluate the intrinsic capabilities of LLMs trained with retrieval augmentation, using a controlled setting. The authors propose an "ideal retrieval" scenario where retrieval is simulated using paraphrases, enabling a cleaner analysis by removing the variability that comes with different retrieval mechanisms or databases. This approach allows for an examination of the impact of pure retrieval augmentation on language processing, independent of the quality of the retrieval data. The models tested include variations with different levels of retrieval noise (0%, 25%, 50%) to simulate varying levels of retrieval quality.

Findings

World Knowledge

Models trained with retrieval augmentation demonstrated lowered performance in tasks related to world knowledge, such as cloze tests from LAMA, indicating these models store less world factual information in their weights. The degradation was more significant as the retrieval noise decreased, suggesting an inverse relationship between retrieval reliance and onboard world knowledge retention.

Syntactic Knowledge

In contrast, syntactic understanding showed consistent improvement across models trained with retrieval augmentation. This enhancement in syntactical tasks indicates that the parameter space within the model, which would otherwise accommodate world knowledge, may be reallocating for better syntactic processing.

Language Understanding

The evaluation pointed to a decline in broader NLU capabilities, especially in tasks requiring the comprehension of extended contexts such as in GLUE and LAMBADA benchmarks. This decline suggests that while retrieval augmentation can offload some memory requirements to external databases, doing so may impair the model's ability to integrate and reason over longer texts internally.

Implications and Future Directions

The observed trade-off between world knowledge retention and syntactic processing efficiency raises critical considerations for the design of retrieval-augmented systems, particularly for applications requiring robust comprehension over extended contexts. The results suggest that while retrieval augmentation can optimize models for specific functionalities, such as syntax parsing, it may not be suitable for tasks requiring extensive internal reasoning and knowledge integration.

Future research could extend these findings by exploring different configurations of retrieval-augmented systems and their impacts on a broader range of linguistic and cognitive capabilities in LLMs. Additionally, studies could investigate the scaling effects of these models to understand how these dynamics play out in larger, more complex systems.

Practical and Theoretical Contributions

From a practical standpoint, these insights could guide the development of more specialized LLMs that either focus on efficient syntactic processing or comprehensive world-knowledge retention based on the needs of the application. Theoretically, this work contributes to our understanding of how external memory aids, such as retrieval systems, interact with the intrinsic learning capabilities of neural models, potentially paving the way for more modular and adaptable AI systems.

File Document Download Save Streamline Icon: https://streamlinehq.com PDF File Document Download Save Streamline Icon: https://streamlinehq.com Markdown Bookmark Chat Bubble Oval Streamline Icon: https://streamlinehq.com Chat (Pro)
Definition Search Book Streamline Icon: https://streamlinehq.com
References (66)
  1. Guillaume Alain and Yoshua Bengio. 2017. Understanding intermediate layers using linear classifier probes. In 5th International Conference on Learning Representations, ICLR 2017, Toulon, France, April 24-26, 2017, Workshop Track Proceedings. OpenReview.net.
  2. The second PASCAL recognising textual entailment challenge. Proceedings of the Second PASCAL Challenges Workshop on Recognising Textual Entailment.
  3. Yonatan Belinkov. 2022. Probing classifiers: Promises, shortcomings, and advances. Computational Linguistics, 48(1):207–219.
  4. The fifth PASCAL recognizing textual entailment challenge. In In Proc Text Analysis Conference (TAC’09.
  5. Improving language models by retrieving from trillions of tokens. In International conference on machine learning, pages 2206–2240. PMLR.
  6. Otakar Borůvka. 1926. O jistém problému minimálním (About a certain minimal problem). Práce moravské přírodovědecké společnosti 3 (1926), 37-58 (1926).
  7. SemEval-2017 task 1: Semantic textual similarity multilingual and crosslingual focused evaluation. In Proceedings of the 11th International Workshop on Semantic Evaluation (SemEval-2017), pages 1–14, Vancouver, Canada. Association for Computational Linguistics.
  8. BRENT: Bidirectional retrieval enhanced Norwegian transformer. In Proceedings of the 24th Nordic Conference on Computational Linguistics (NoDaLiDa), pages 202–214, Tórshavn, Faroe Islands. University of Tartu Library.
  9. Yeong-Jin Chu and Tseng-Hong Liu. 1965. On the shortest arborescence of a directed graph. Science Sinica, 14:1396–1400.
  10. The PASCAL recognising textual entailment challenge. In Machine Learning Challenges. Evaluating Predictive Uncertainty, Visual Object Classification, and Recognising Tectual Entailment, pages 177–190, Berlin, Heidelberg. Springer Berlin Heidelberg.
  11. Editing factual knowledge in language models. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 6491–6506, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics.
  12. An optimal transportation approach for assessing almost stochastic order. In The Mathematics of the Uncertain, pages 33–44. Springer.
  13. BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 4171–4186, Minneapolis, Minnesota. Association for Computational Linguistics.
  14. William B. Dolan and Chris Brockett. 2005. Automatically constructing a corpus of sentential paraphrases. In Proceedings of the Third International Workshop on Paraphrasing (IWP2005).
  15. Surface-based retrieval reduces perplexity of retrieval-augmented language models.
  16. Timothy Dozat and Christopher D. Manning. 2017. Deep biaffine attention for neural dependency parsing. In International Conference on Learning Representations.
  17. Deep dominance - how to properly compare deep neural models. In Proceedings of the 57th Conference of the Association for Computational Linguistics, ACL 2019, Florence, Italy, July 28- August 2, 2019, Volume 1: Long Papers, pages 2773–2785. Association for Computational Linguistics.
  18. T-REx: A large scale alignment of natural language with knowledge base triples. In Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018), Miyazaki, Japan. European Language Resources Association (ELRA).
  19. The third PASCAL recognizing textual entailment challenge. In Proceedings of the ACL-PASCAL Workshop on Textual Entailment and Paraphrasing, pages 1–9, Prague. Association for Computational Linguistics.
  20. Retrieval augmented language model pre-training. In Proceedings of the 37th International Conference on Machine Learning, volume 119 of Proceedings of Machine Learning Research, pages 3929–3938. PMLR.
  21. The curious case of neural text degeneration. In International Conference on Learning Representations.
  22. Do attention heads in bert track syntactic dependencies? NY Academy of Sciences NLP, Dialog, and Speech Workshop.
  23. Atlas: Few-shot learning with retrieval augmented language models. Journal of Machine Learning Research, 24(251):1–43.
  24. Mistral 7B.
  25. Dan Kondratyuk and Milan Straka. 2019. 75 languages, 1 model: Parsing Universal Dependencies universally. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 2779–2795, Hong Kong, China. Association for Computational Linguistics.
  26. The Winograd schema challenge. In AAAI Spring Symposium: Logical Formalizations of Commonsense Reasoning, volume 46, page 47.
  27. Boosted dense retriever. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 3102–3117, Seattle, United States. Association for Computational Linguistics.
  28. Linguistic knowledge and transferability of contextual representations. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 1073–1094, Minneapolis, Minnesota. Association for Computational Linguistics.
  29. David Mareček and Rudolf Rosa. 2018. Extracting syntactic trees from transformer encoder self-attentions. In Proceedings of the 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP, pages 347–349, Brussels, Belgium. Association for Computational Linguistics.
  30. B.W. Matthews. 1975. Comparison of the predicted and observed secondary structure of t4 phage lysozyme. Biochimica et Biophysica Acta (BBA) - Protein Structure, 405(2):442–451.
  31. Unsupervised paraphrasing with pretrained language models. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 5136–5150, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics.
  32. Joakim Nivre and Chiao-Ting Fang. 2017. Universal Dependency evaluation. In Proceedings of the NoDaLiDa 2017 Workshop on Universal Dependencies (UDW 2017), pages 86–95, Gothenburg, Sweden. Association for Computational Linguistics.
  33. Universal Dependencies. In Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Tutorial Abstracts, Valencia, Spain. Association for Computational Linguistics.
  34. On the generalization ability of retrieval-enhanced transformers. In Findings of the Association for Computational Linguistics: EACL 2023, pages 1485–1493, Dubrovnik, Croatia. Association for Computational Linguistics.
  35. The LAMBADA dataset: Word prediction requiring a broad discourse context. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1525–1534, Berlin, Germany. Association for Computational Linguistics.
  36. Bleu: a method for automatic evaluation of machine translation. In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics, pages 311–318, Philadelphia, Pennsylvania, USA. Association for Computational Linguistics.
  37. Language models as knowledge bases? In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 2463–2473, Hong Kong, China. Association for Computational Linguistics.
  38. Matt Post. 2018. A call for clarity in reporting BLEU scores. In Proceedings of the Third Conference on Machine Translation: Research Papers, pages 186–191, Brussels, Belgium. Association for Computational Linguistics.
  39. Language models are unsupervised multitask learners. OpenAI blog, 1(8):9.
  40. Alessandro Raganato and Jörg Tiedemann. 2018. An analysis of encoder representations in transformer-based machine translation. In Proceedings of the 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP, pages 287–297, Brussels, Belgium. Association for Computational Linguistics.
  41. SQuAD: 100,000+ questions for machine comprehension of text. In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pages 2383–2392, Austin, Texas. Association for Computational Linguistics.
  42. Attention can reflect syntactic structure (if you let it). In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, pages 3031–3045, Online. Association for Computational Linguistics.
  43. Nils Reimers and Iryna Gurevych. 2019. Sentence-BERT: Sentence embeddings using Siamese BERT-networks. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 3982–3992, Hong Kong, China. Association for Computational Linguistics.
  44. A primer in BERTology: What we know about how BERT works. Transactions of the Association for Computational Linguistics, 8:842–866.
  45. Masked language model scoring. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 2699–2712, Online. Association for Computational Linguistics.
  46. David Samuel. 2023. Mean BERTs make erratic language teachers: the effectiveness of latent bootstrapping in low-resource settings. In Proceedings of the BabyLM Challenge at the 27th Conference on Computational Natural Language Learning, pages 193–209, Singapore. Association for Computational Linguistics.
  47. Trained on 100 million words and still in shape: BERT meets British National Corpus. In Findings of the Association for Computational Linguistics: EACL 2023, pages 1954–1974, Dubrovnik, Croatia. Association for Computational Linguistics.
  48. Does string-based neural MT learn source syntax? In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pages 1526–1534, Austin, Texas. Association for Computational Linguistics.
  49. A gold standard dependency corpus for English. In Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC’14), pages 2897–2904, Reykjavik, Iceland. European Language Resources Association (ELRA).
  50. Recursive deep models for semantic compositionality over a sentiment treebank. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 1631–1642, Seattle, Washington, USA. Association for Computational Linguistics.
  51. Conceptnet 5.5: An open multilingual graph of general knowledge. In Proceedings of the AAAI conference on artificial intelligence, volume 31.
  52. Llama 2: Open foundation and fine-tuned chat models.
  53. deep-significance: Easy and meaningful signifcance testing in the age of neural networks. In ML Evaluation Standards Workshop at the Tenth International Conference on Learning Representations.
  54. Attention is all you need. In Advances in Neural Information Processing Systems, volume 30. Curran Associates, Inc.
  55. Alex Wang and Kyunghyun Cho. 2019. BERT has a mouth, and it must speak: BERT as a Markov random field language model. In Proceedings of the Workshop on Methods for Optimizing and Evaluating Neural Language Generation, pages 30–36, Minneapolis, Minnesota. Association for Computational Linguistics.
  56. GLUE: A multi-task benchmark and analysis platform for natural language understanding. In Proceedings of the 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP, pages 353–355, Brussels, Belgium. Association for Computational Linguistics.
  57. Shall we pretrain autoregressive language models with retrieval? a comprehensive study. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 7763–7786, Singapore. Association for Computational Linguistics.
  58. k𝑘kitalic_kNN-LM does not improve open-ended text generation. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 15023–15037, Singapore. Association for Computational Linguistics.
  59. BLiMP: The benchmark of linguistic minimal pairs for English. Transactions of the Association for Computational Linguistics, 8:377–392.
  60. Neural network acceptability judgments. Transactions of the Association for Computational Linguistics, 7:625–641.
  61. Learning which features matter: RoBERTa acquires a preference for linguistic generalizations (eventually). In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 217–235, Online. Association for Computational Linguistics.
  62. A broad-coverage challenge corpus for sentence understanding through inference. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), pages 1112–1122, New Orleans, Louisiana. Association for Computational Linguistics.
  63. Colin Wilson. 2006. Learning phonology with substantive bias: An experimental and computational study of velar palatalization. Cognitive Science, 30(5):945–982.
  64. Transformers: State-of-the-art natural language processing. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, pages 38–45, Online. Association for Computational Linguistics.
  65. Google’s neural machine translation system: Bridging the gap between human and machine translation.
  66. Editing large language models: Problems, methods, and opportunities.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (3)
  1. David Samuel (23 papers)
  2. Lucas Georges Gabriel Charpentier (8 papers)
  3. Sondre Wold (9 papers)
Citations (1)
View on Semantic Scholar