Elamur Research⌕
☰
home / papers / 2023 / Rethinking with Retrieval: Faithful Large Language Model Inference

Rethinking with Retrieval: Faithful Large Language Model Inference

Paper · 2023 · Hangfeng He, Hongming Zhang and Dan Roth

Summary

Large language models (LLMs) often hold knowledge that is missing, stale or wrong. Most ways of adding external facts require extra training, which is impractical at GPT-3 scale. He, Zhang and Roth propose rethinking with retrieval (RR), a post-processing step that needs no fine-tuning and is not limited by input length. The model first samples several chain-of-thought (CoT) reasoning paths. Each individual reasoning step then serves as a query to fetch evidence from an external source: Wikipedia, Wikidata, WordNet or ConceptNet, depending on the task. Each path is scored for how well the evidence supports it, and the answer with the highest total score is chosen. Two extensions either pick supported sentences from different paths or rewrite unsupported claims using retrieved text. UnifiedQA-v2 then derives the final answer from these facts. With GPT-3 175B on commonsense, temporal and tabular reasoning, RR beat zero/few-shot prompting, CoT and self-consistency on every task. It also produced explanations better grounded in evidence. The work shows that retrieval can make LLM reasoning more reliable without changing model weights.

Key points

  • On StrategyQA with the dataset's evidence paragraphs, CoT prompting reached 65.94% accuracy and 38.73% faithfulness. Basic RR without voting reached 76.86% / 50.02%. Variant I (fact selection) reached 78.60% / 54.11%, and Variant II (fact generation) reached 78.60% / 54.54%.
  • With voting, basic RR scored 79.91% using evidence paragraphs and 77.73% using retrieved paragraphs. Both exceed self-consistency at 73.36%.
  • Faithfulness was measured with an NLI-based factual consistency metric (following Honovich et al., 2022), using the pre-trained NLI model of Nie et al. (2020).
  • UnifiedQA-v2 alone scored only 58.95% on StrategyQA but 90.83% when given gold supporting facts. It was therefore used only for the final inference step.
  • Variant II outperformed CoT in both accuracy and faithfulness across OPT models from 1.3B to 30B and GPT-3 175B. Code is available at github.com/HornHehhf/RR.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓University of PennsylvaniaPhiladelphia · United States233 works · 65,583 citations🎓University of RochesterRochester · United States43 works · 7,874 citations

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 worksComputer Vision and Pattern Recognition · 3,486 works

🧭 Topics 2Research subjects and languages the work is about.Language models & pretraining · 3,894 worksInformation retrieval & search · 552 works

Cite

Cite as Hangfeng He, Hongming Zhang and Dan Roth. 2023. Rethinking with Retrieval: Faithful Large Language Model Inference. arXiv:2301.00303.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.2301.00303 · arXiv 2301.00303

@misc{he2023rethinking,
  title = {{Rethinking with Retrieval: Faithful Large Language Model Inference}},
  author = {Hangfeng He, Hongming Zhang and Dan Roth},
  year = {2023},
  doi = {10.48550/arxiv.2301.00303},
  eprint = {2301.00303},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2301.00303}
}

References

53 references, 35 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Kurt Bollacker, Colin Evans, Praveen Paritosh, Tim Sturge, and Jamie Taylor. 2008. Freebase: a collaboratively created graph database for structuring human knowledge. In Proceedings of the 2008 ACM SIGMOD international conference on Management of data, pages 1247– 1250. find
  2. Sebastian Borgeaud, Arthur Mensch, Jordan Hoffmann, Trevor Cai, Eliza Rutherford, Katie Millican, George van den Driessche, Jean-Baptiste Lespiau, Bogdan Damoc, Aidan Clark, et al. 2021. Improving language models by retrieving from trillions of tokens. arXiv preprint arXiv:2112.04426. find
  3. Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. 2020. Language models are few-shot learners. Advances in neural information processing systems, 33:1877–1901.
  4. Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, Maarten Bosma, Gaurav Mishra, Adam Roberts, Paul Barham, Hyung Won Chung, Charles Sutton, Sebastian Gehrmann, et al. 2022. Palm: Scaling language modeling with pathways. arXiv preprint arXiv:2204.02311.
  5. Karl Cobbe, Vineet Kosaraju, Mohammad Bavarian, Jacob Hilton, Reiichiro Nakano, Christopher Hesse, and John Schulman. 2021. Training verifiers to solve math word problems. arXiv preprint arXiv:2110.14168.
  6. Ido Dagan, Oren Glickman, and Bernardo Magnini. 2005. The pascal recognising textual entailment challenge. In Machine learning challenges workshop, pages 177–190. Springer. find
  7. Daniel Deutsch, Tania Bedrax-Weiss, and Dan Roth. 2021. Towards question-answering as an automatic metric for evaluating the content quality of a summary. Transactions of the Association for Computational Linguistics, 9:774– 789.
  8. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019
  9. Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 4171–4186. find
  10. Alexander R Fabbri, Chien-Sheng Wu, Wenhao Liu, and Caiming Xiong. 2021. Qafacteval: Improved qa-based factual consistency evaluation for summarization. arXiv preprint arXiv:2112.08542.
  11. Mor Geva, Daniel Khashabi, Elad Segal, Tushar Khot, Dan Roth, and Jonathan Berant. 2021. Did aristotle use a laptop? a question answering benchmark with implicit reasoning strategies. Transactions of the Association for Computational Linguistics, 9:346–361. find
  12. Liangke Gui, Borui Wang, Qiuyuan Huang, Alex Hauptmann, Yonatan Bisk, and Jianfeng Gao. 2021. Kat: A knowledge augmented transformer for vision-and-language. arXiv preprint arXiv:2112.08614. find
  13. Vivek Gupta, Maitrey Mehta, Pegah Nokhiz, and Vivek Srikumar. 2020. Infotabs: Inference on tables as semi-structured data. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 2309–2324.
  14. Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat, and Mingwei Chang. 2020. Retrieval augmented language model pre-training. In International Conference on Machine Learning, pages 3929–3938. PMLR.
  15. Or Honovich, Roee Aharoni, Jonathan Herzig, Hagai Taitelbaum, Doron Kukliansy, Vered Cohen, Thomas Scialom, Idan Szpektor, Avinatan Hassidim, and Yossi Matias. 2022. True: Re-evaluating factual consistency evaluation. In Proceedings of the Second DialDoc Workshop on Document-grounded Dialogue and Co…
  16. Or Honovich, Leshem Choshen, Roee Aharoni, Ella Neeman, Idan Szpektor, and Omri Abend. 2021. Q2:: Evaluating factual consistency in knowledge-grounded dialogues via question generation and question answering. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing,… find
  17. Zhen Jia, Abdalghani Abujabal, Rishiraj Saha Roy, Jannik Strötgen, and Gerhard Weikum. 2018. Tempquestions: A benchmark for temporal question answering. In Companion Proceedings of the The Web Conference 2018, pages 1057–1062. find
  18. Mandar Joshi, Kenton Lee, Yi Luan, and Kristina Toutanova. 2020. Contextualized representations using textual encyclopedic knowledge. arXiv preprint arXiv:2004.12006. — [arXiv:2004.12006](https://arxiv.org/abs/2004.12006) find
  19. Vladimir Karpukhin, Barlas Oguz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih. 2020. Dense passage retrieval for open-domain question answering. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 6769–6781.
  20. Urvashi Khandelwal, Omer Levy, Dan Jurafsky, Luke Zettlemoyer, and Mike Lewis. 2020. Generalization through memorization: Nearest neighbor language models. In International Conference on Learning Representations. find
  21. Daniel Khashabi, Yeganeh Kordi, and Hannaneh Hajishirzi. 2022. Unifiedqa-v2: Stronger generalization via broader cross-format training. arXiv preprint arXiv:2202.12359. — [arXiv:2202.12359](https://arxiv.org/abs/2202.12359) find
  22. Daniel Khashabi, Sewon Min, Tushar Khot, Ashish Sabharwal, Oyvind Tafjord, Peter Clark, and Hannaneh Hajishirzi. 2020. Unifiedqa: Crossing format boundaries with a single qa system. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 1896–1907.
  23. Takeshi Kojima, Shixiang Shane Gu, Machel Reid, Yutaka Matsuo, and Yusuke Iwasawa. 2022. Large language models are zero-shot reasoners. arXiv preprint arXiv:2205.11916.
  24. Mojtaba Komeili, Kurt Shuster, and Jason Weston. 2022. Internet-augmented dialogue generation. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 8460–8478.
  25. Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, et al. 2020. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in Neural Information Processing Systems, 33:9459–9474.
  26. Jimmy Lin, Xueguang Ma, Sheng-Chieh Lin, Jheng-Hong Yang, Ronak Pradeep, and Rodrigo Nogueira. 2021. Pyserini: A Python toolkit for reproducible information retrieval research with sparse and dense representations. In Proceedings of the 44th Annual International ACM SI-GIR Conference on Research and… find
  27. Jiacheng Liu, Alisa Liu, Ximing Lu, Sean Welleck, Peter West, Ronan Le Bras, Yejin Choi, and Hannaneh Hajishirzi. 2022. Generated knowledge prompting for commonsense reasoning. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 3… find
  28. Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692.
  29. George A Miller. 1995. Wordnet: a lexical database for english. Communications of the ACM, 38(11):39–41.
  30. Reiichiro Nakano, Jacob Hilton, Suchir Balaji, Jeff Wu, Long Ouyang, Christina Kim, Christopher Hesse, Shantanu Jain, Vineet Kosaraju, William Saunders, et al. 2021. Webgpt: Browser-assisted question-answering with human feedback. arXiv preprint arXiv:2112.09332.
  31. J Neeraja, Vivek Gupta, and Vivek Srikumar. 2021. Incorporating external knowledge to enhance tabular reasoning. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 2799–2809. find
  32. Yixin Nie, Adina Williams, Emily Dinan, Mohit Bansal, Jason Weston, and Douwe Kiela. 2020. Adversarial nli: A new benchmark for natural language understanding. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4885–4901.
  33. Maxwell Nye, Anders Johan Andreassen, Guy Gur-Ari, Henryk Michalewski, Jacob Austin, David Bieber, David Dohan, Aitor Lewkowycz, Maarten Bosma, David Luan, et al. 2022. Show your work: Scratchpads for intermediate computation with language models. In Deep Learning for Code Workshop.
  34. Maxwell Nye, Michael Tessler, Josh Tenenbaum, and Brenden M Lake. 2021. Improving coherence and consistency in neural sequence models with dual-system, neuro-symbolic reasoning. Advances in Neural Information Processing Systems, 34:25192–25204. find
  35. Long Ouyang, Jeff Wu, Xu Jiang, Diogo Almeida, Carroll L Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, et al. 2022. Training language models to follow instructions with human feedback. arXiv preprint arXiv:2203.02155.
  36. Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J Liu. 2020. Exploring the limits of transfer learning with a unified text-to-text transformer. Journal of Machine Learning Research, 21:1–67.
  37. Pranav Rajpurkar, Jian Zhang, Konstantin Lopyrev, and Percy Liang. 2016. 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.
  38. Stephen Robertson, Hugo Zaragoza, et al. 2009. The probabilistic relevance framework: Bm25 and beyond. Foundations and Trends® in Information Retrieval, 3(4):333–389. find
  39. Kurt Shuster, Mojtaba Komeili, Leonard Adolphs, Stephen Roller, Arthur Szlam, and Jason Weston. 2022. Language models that seek for knowledge: Modular search & generation for dialogue and prompt completion. arXiv preprint arXiv:2203.13224.
  40. Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, and Tie-Yan Liu. 2020. Mpnet: Masked and permuted pre-training for language understanding. Advances in Neural Information Processing Systems, 33:16857–16867.
  41. Robyn Speer, Joshua Chin, and Catherine Havasi. 2017. Conceptnet 5.5: An open multilingual graph of general knowledge. In Thirty-first AAAI conference on artificial intelligence. find
  42. Alon Talmor, Oyvind Tafjord, Peter Clark, Yoav Goldberg, and Jonathan Berant. 2020. Leap-of-thought: Teaching pre-trained models to systematically reason over implicit knowledge. Advances in Neural Information Processing Systems, 33:20227–20237.
  43. Romal Thoppilan, Daniel De Freitas, Jamie Hall, Noam Shazeer, Apoorv Kulshreshtha, Heng-Tze Cheng, Alicia Jin, Taylor Bos, Leslie Baker, Yu Du, et al. 2022. Lamda: Language models for dialog applications. arXiv preprint arXiv:2201.08239.
  44. Yerram Varun, Aayush Sharma, and Vivek Gupta. 2022. Trans-kblstm: An external knowledge enhanced transformer bilstm model for tabular reasoning. In Proceedings of Deep Learning Inside Out (DeeLIO 2022): The 3rd Workshop on Knowledge Extraction and Integration for Deep Learning Architectures, pages 6… find
  45. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. Advances in neural information processing systems, 30.
  46. Denny Vrandečić and Markus Krötzsch. 2014. Wikidata: a free collaborative knowledgebase. Communications of the ACM, 57(10):78–85.
  47. Xuezhi Wang, Jason Wei, Dale Schuurmans, Quoc Le, Ed Chi, and Denny Zhou. 2022. Self-consistency improves chain of thought reasoning in language models. arXiv preprint arXiv:2203.11171.
  48. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Ed Chi, Quoc Le, and Denny Zhou. 2022. Chain of thought prompting elicits reasoning in large language models. arXiv preprint arXiv:2201.11903.
  49. Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pierric Cistac, Tim Rault, Rémi Louf, Morgan Funtowicz, et al. 2020. Transformers: State-of-the-art natural language processing. In Proceedings of the 2020 conference on empirical methods in natural language pr…
  50. Xi Ye and Greg Durrett. 2022. The unreliability of explanations in few-shot in-context learning. arXiv preprint arXiv:2205.03401.
  51. Eric Zelikman, Yuhuai Wu, and Noah D Goodman. 2022. Star: Bootstrapping reasoning with reasoning. arXiv preprint arXiv:2203.14465.
  52. Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen, Christopher Dewan, Mona Diab, Xian Li, Xi Victoria Lin, et al. 2022. Opt: Open pre-trained transformer language models. arXiv preprint arXiv:2205.01068.
  53. Denny Zhou, Nathanael Schärli, Le Hou, Jason Wei, Nathan Scales, Xuezhi Wang, Dale Schuurmans, Olivier Bousquet, Quoc Le, and Ed Chi. 2022. Least-to-most prompting enables complex reasoning in large language models. arXiv preprint arXiv:2205.10625.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/he-2023-rethinking-retrieval-faithful-large-language/