Elamur Research
☰
home / papers / 2023 / Enabling Large Language Models to Generate Text with Citations

Enabling Large Language Models to Generate Text with Citations

Written for the Research library (claude-opus-5-5) from the paper text. Original — aclanthology.org/2023.emnlp-main.398.pdf

Summary

LLMs used for information seeking often hallucinate, and their answers are hard to check without sources. Earlier work on making models cite evidence relied on commercial search engines, closed models and costly human judging, so its results were hard to reproduce or compare. The authors introduce ALCE, a reproducible benchmark for automatically evaluating LLM answers with citations. A system gets a question and a large corpus. It must retrieve passages, write a long answer, and attach passage citations to each sentence. ALCE draws on three QA datasets: ASQA, QAMPARI and ELI5. Answers are scored on three axes: fluency (MAUVE), correctness (metrics tailored to each dataset) and citation quality (citation recall and precision, judged by an NLI model). Human evaluation agrees well with these automatic scores. Tests with ChatGPT, GPT-4, Vicuna and several prompting strategies show large gaps. On ELI5, about half of the best models' answers are not fully backed by the passages they cite. The benchmark makes progress on verifiable, retrieval-grounded generation measurable.

Key points

  • Data: 1,000 development examples from each of ASQA (ambiguous factoid questions), QAMPARI (list-answer questions) and ELI5 (how/why questions from Reddit). The corpora are split into 100-word passages: a 2018-12-20 Wikipedia snapshot (21M passages) for the first two datasets and Sphere, a filtered Common Crawl (899M passages), for ELI5. Each statement can cite at most 3 passages.
  • Metrics: fluency uses MAUVE, mainly as a sanity check. Correctness uses EM recall of short answers for ASQA and precision plus recall-5 for QAMPARI, where finding 5 correct entities counts as full recall. Citation recall and precision come from an NLI entailment model.
  • Validation against humans: Cohen's kappa is 0.698 for citation recall and 0.525 for citation precision. Taking human labels as gold, the automatic metrics are 85.1% and 77.6% accurate.
  • Findings:
    • A closed-book model that adds citations afterwards gets answers right but cites poorly.
    • Interactive retrieval does not help on this benchmark.
    • Summarizing retrieved passages improves correctness but not citations.
    • Reranking several generations improves citation quality.
    • Adding more passages to the context helps GPT-4 but not ChatGPT.
  • Open challenges and resources: the authors name weak retrieval, limited context length and difficulty combining several sources while ignoring irrelevant ones. Code and data are available at github.com/princeton-nlp/ALCE.

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

Fields: Natural language processing

Cited by

273 papers of our library cite it (of the 140673 whose bibliographies we read); 10 of them have a page here, 263 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2023.emnlp-main.398

@inproceedings{gao2023enabling,
  title = {{Enabling Large Language Models to Generate Text with Citations}},
  author = {Gao, Tianyu and Yen, Howard and Yu, Jiatong and Chen, Danqi},
  booktitle = {Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing},
  year = {2023},
  doi = {10.18653/v1/2023.emnlp-main.398},
  url = {https://doi.org/10.18653/v1/2023.emnlp-main.398}
}

References

55 references, 27 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Bernd Bohnet, Vinh Q Tran, Pat Verga, Roee Aharoni, Daniel Andor, Livio Baldini Soares, Jacob Eisenstein, Kuzman Ganchev, Jonathan Herzig, Kai Hui, et al. 2022. Attributed question answering: Evaluation and modeling for attributed large language models. arXiv preprint arXiv:2212.08037. find our card
  2. Sebastian Borgeaud, Arthur Mensch, Jordan Hoffmann, Trevor Cai, Eliza Rutherford, Katie Millican, George Bm Van Den Driessche, Jean-Baptiste Lespiau, Bogdan Damoc, Aidan Clark, Diego De Las Casas, Aurelia Guy, Jacob Menick, Roman Ring, Tom Hennigan, Saffron Huang, Loren Maggiore, Chris Jones, Albin… find
  3. Samuel Bowman, Gabor Angeli, Christopher Potts, and Christopher D Manning. 2015. A large annotated corpus for learning natural language inference. In Empirical Methods in Natural Language Processing (EMNLP). find
  4. Tom B Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. 2020. Language models are few-shot learners. In Advances in Neural Information Processing Systems (NeurIPS). find
  5. Wei-Lin Chiang, Zhuohan Li, Zi Lin, Ying Sheng, Zhanghao Wu, Hao Zhang, Lianmin Zheng, Siyuan Zhuang, Yonghao Zhuang, Joseph E. Gonzalez, Ion Stoica, and Eric P. Xing. 2023. Vicuna: An open-source chatbot impressing gpt-4 with 90%* chatgpt quality. find
  6. Angela Fan, Yacine Jernite, Ethan Perez, David Grangier, Jason Weston, and Michael Auli. 2019. ELI5: Long form question answering. In Association for Computational Linguistics (ACL), pages 3558–3567. find our card
  7. Martin Funkquist, Ilia Kuznetsov, Yufang Hou, and Iryna Gurevych. 2022. CiteBench: A benchmark for Scientific Citation Text Generation. arXiv preprint arXiv:2212.09577. find
  8. Luyu Gao, Zhuyun Dai, Panupong Pasupat, Anthony Chen, Arun Tejasvi Chaganty, Yicheng Fan, Vincent Y Zhao, Ni Lao, Hongrae Lee, Da-Cheng Juan, et al. 2023. RARR: Researching and revising what language models say, using language models. In Association for Computational Linguistics (ACL). find our card
  9. Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat, and Ming-Wei Chang. 2020. REALM: Retrieval-augmented language model pre-training. In International Conference on Machine Learning (ICML). find
  10. Hangfeng He, Hongming Zhang, and Dan Roth. 2022. Rethinking with retrieval: Faithful large language model inference. arXiv preprint arXiv:2301.00303. find
  11. Ari Holtzman, Jan Buys, Li Du, Maxwell Forbes, and Yejin Choi. 2020. The curious case of neural text degeneration. In International Conference on Learning Representations (ICLR). find
  12. 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 North American Chapter of the Association for Computational Linguistics: Human… find our card
  13. Gautier Izacard and Edouard Grave. 2021. Leveraging passage retrieval with generative models for open domain question answering. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, pages 874–880, Online. Association for Computa… find our card
  14. Gautier Izacard, Patrick Lewis, Maria Lomeli, Lucas Hosseini, Fabio Petroni, Timo Schick, Jane Dwivedi-Yu, Armand Joulin, Sebastian Riedel, and Edouard Grave. 2022. Atlas: Few-shot learning with retrieval augmented language models. arXiv preprint arXiv:2208.03299. find
  15. Ziwei Ji, Nayeon Lee, Rita Frieske, Tiezheng Yu, Dan Su, Yan Xu, Etsuko Ishii, Ye Jin Bang, Andrea Madotto, and Pascale Fung. 2023. Survey of hallucination in natural language generation. ACM Computing Surveys, 55(12):1–38. find our card
  16. Zhengbao Jiang, Frank F Xu, Luyu Gao, Zhiqing Sun, Qian Liu, Jane Dwivedi-Yu, Yiming Yang, Jamie Callan, and Graham Neubig. 2023. Active retrieval augmented generation. arXiv preprint arXiv:2305.06983. find
  17. Ryo Kamoi, Tanya Goyal, Juan Diego Rodriguez, and Greg Durrett. 2023. WiCE: Real-World Entailment for Claims in Wikipedia. arXiv preprint arXiv:2303.01432. find
  18. 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, Onli… find our card
  19. 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 (ICLR). find
  20. Tushar Khot, Ashish Sabharwal, and Peter Clark. 2018. Scitail: A textual entailment dataset from science question answering. In Conference on Artificial Intelligence (AAAI), volume 32. find
  21. Kalpesh Krishna, Aurko Roy, and Mohit Iyyer. 2021. Hurdles to progress in long-form question answering. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 4940–4957, Online. Association for Computat… find our card
  22. Andreas Köpf, Yannic Kilcher, Dimitri von Rütte, Sotiris Anagnostidis, Zhi-Rui Tam, Keith Stevens, Abdullah Barhoum, Nguyen Minh Duc, Oliver Stanley, Richárd Nagyfi, Shahul ES, Sameer Suri, David Glushkov, Arnav Dantuluri, Andrew Maguire, Christoph Schuhmann, Huu Nguyen, and Alexander Mattick. 2023.… find our card
  23. Kenton Lee, Ming-Wei Chang, and Kristina Toutanova. 2019. Latent retrieval for weakly supervised open domain question answering. In Association for Computational Linguistics (ACL), pages 6086–6096. find our card
  24. Nelson F Liu, Tianyi Zhang, and Percy Liang. 2023. Evaluating verifiability in generative search engines. arXiv preprint arXiv:2304.09848. find our card
  25. Jacob Menick, Maja Trebacz, Vladimir Mikulik, John Aslanides, Francis Song, Martin Chadwick, Mia Glaese, Susannah Young, Lucy Campbell-Gillingham, Geoffrey Irving, et al. 2022. Teaching language models to support answers with verified quotes. arXiv preprint arXiv:2203.11147. find our card
  26. Sewon Min, Julian Michael, Hannaneh Hajishirzi, and Luke Zettlemoyer. 2020. AmbigQA: Answering ambiguous open-domain questions. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 5783– 5797, Online. Association for Computational Linguistics. find
  27. 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. find our card
  28. Jianmo Ni, Chen Qu, Jing Lu, Zhuyun Dai, Gustavo Hernandez Abrego, Ji Ma, Vincent Zhao, Yi Luan, Keith Hall, Ming-Wei Chang, and Yinfei Yang. 2022. Large dual encoders are generalizable retrievers. In Empirical Methods in Natural Language Processing (EMNLP), pages 9844–9855. find our card
  29. OpenAI. 2023. GPT-4 Technical Report. find our card
  30. Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, et al. 2022. Training language models to follow instructions with human feedback. Advances in Neural Information Processing Systems (NeurIPS), 35:27730– 2774… find
  31. Aaron Parisi, Yao Zhao, and Noah Fiedel. 2022. TALM: Tool augmented language models. arXiv preprint arXiv:2205.12255. find our card
  32. Fabio Petroni, Aleksandra Piktus, Angela Fan, Patrick Lewis, Majid Yazdani, Nicola De Cao, James Thorne, Yacine Jernite, Vladimir Karpukhin, Jean Maillard, Vassilis Plachouras, Tim Rocktäschel, and Sebastian Riedel. 2021. KILT: a benchmark for knowledge intensive language tasks. In Proceedings of th… find our card
  33. Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Dmytro Okhonko, Samuel Broscheit, Gautier Izacard, Patrick Lewis, Barlas Oğuz, Edouard Grave, Wen-tau Yih, et al. 2021. The Web Is Your Oyster– Knowledge-Intensive NLP against a Very Large Web Corpus. arXiv preprint arXiv:2112.09924. — [arXiv:2112.09924](https://arxiv.org/abs/2112.09924) find
  34. Krishna Pillutla, Swabha Swayamdipta, Rowan Zellers, John Thickstun, Sean Welleck, Yejin Choi, and Zaid Harchaoui. 2021. MAUVE: Measuring the gap between neural text and human text using divergence frontiers. In Advances in Neural Information Processing Systems. find our card
  35. Ofir Press, Muru Zhang, Sewon Min, Ludwig Schmidt, Noah A Smith, and Mike Lewis. 2022. Measuring and narrowing the compositionality gap in language models. arXiv preprint arXiv:2210.03350. find
  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. The Journal of Machine Learning Research (JMLR), 21(140). find
  37. Pranav Rajpurkar, Jian Zhang, Konstantin Lopyrev, and Percy Liang. 2016. SQuAD: 100,000+ questions for machine comprehension of text. In Empirical Methods in Natural Language Processing (EMNLP), pages 2383–2392. find our card
  38. Hannah Rashkin, Vitaly Nikolaev, Matthew Lamm, Lora Aroyo, Michael Collins, Dipanjan Das, Slav Petrov, Gaurav Singh Tomar, Iulia Turc, and David Reitter. 2023. Measuring Attribution in Natural Language Generation Models. Computational Linguistics, pages 1–64. find our card
  39. Samuel Joseph Amouyal Ohad Rubin, Ori Yoran, Tomer Wolfson, Jonathan Herzig, and Jonathan Berant. 2022. QAMPARI: An Open-domain Question Answering Benchmark for Questions with Many Answers from Multiple Paragraphs. arXiv preprint arXiv:2205.12665. — [arXiv:2205.12665](https://arxiv.org/abs/2205.12665) find
  40. Timo Schick, Jane Dwivedi-Yu, Roberto Dessì, Roberta Raileanu, Maria Lomeli, Luke Zettlemoyer, Nicola Cancedda, and Thomas Scialom. 2023. Toolformer: Language models can teach themselves to use tools. arXiv preprint arXiv:2302.04761. find our card
  41. Tal Schuster, Adam Fisch, and Regina Barzilay. 2021. Get your vitamin C! robust fact verification with contrastive evidence. In North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pages 624– 643. find our card
  42. Freda Shi, Xinyun Chen, Kanishka Misra, Nathan Scales, David Dohan, Ed Chi, Nathanael Schärli, and Denny Zhou. 2023. Large language models can be easily distracted by irrelevant context. In International Conference on Machine Learning (ICML). find our card
  43. 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. In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 373–393. find
  44. Ivan Stelmakh, Yi Luan, Bhuwan Dhingra, and Ming-Wei Chang. 2022. ASQA: Factoid questions meet long-form answers. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 8273–8288, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics. find our card
  45. Rohan Taori, Ishaan Gulrajani, Tianyi Zhang, Yann Dubois, Xuechen Li, Carlos Guestrin, Percy Liang, and Tatsunori B. Hashimoto. 2023. Stanford Alpaca: An Instruction-following LLaMA model. find
  46. James Thorne, Andreas Vlachos, Christos Christodoulopoulos, and Arpit Mittal. 2018. FEVER: a large-scale dataset for fact extraction and VERification. In North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pages 809–819. find our card
  47. Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, et al. 2023a. LLaMA: Open and Efficient Foundation Language Models. arXiv preprint arXiv:2302.13971. find
  48. Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, Dan Bikel, Lukas Blecher, Cristian Canton Ferrer, Moya Chen, Guillem Cucurull, David Esiobu, Jude Fernandes, Jeremy Fu, Wenyin Fu, Brian Fuller,… find
  49. Alex Wang, Kyunghyun Cho, and Mike Lewis. 2020. Asking and answering questions to evaluate the factual consistency of summaries. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 5008–5020, Online. Association for Computational Linguistics. find
  50. Adina Williams, Nikita Nangia, and Samuel Bowman. 2018. A broad-coverage challenge corpus for sentence understanding through inference. In North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT). find
  51. Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik R Narasimhan, and Yuan Cao. 2023. React: Synergizing reasoning and acting in language models. In International Conference on Learning Representations (ICLR). find our card
  52. Xiang Yue, Boshi Wang, Kai Zhang, Ziru Chen, Yu Su, and Huan Sun. 2023. Automatic evaluation of attribution by large language models. arXiv preprint arXiv:2305.06311. find
  53. Shiyue Zhang and Mohit Bansal. 2021. Finding a balanced degree of automation for summary evaluation. In Empirical Methods in Natural Language Processing (EMNLP), pages 6617–6632. find
  54. Yuan Zhang, Jason Baldridge, and Luheng He. 2019. PAWS: Paraphrase adversaries from word scrambling. In North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pages 1298–1308. find our card
  55. Zexuan Zhong, Tao Lei, and Danqi Chen. 2022. Training language models with memory augmentation. In Empirical Methods in Natural Language Processing (EMNLP), pages 5657–5673. find