Elamur Research⌕
☰
home / papers / 2022 / Ground-Truth Labels Matter: A Deeper Look into Input-Label Demonstrations

Ground-Truth Labels Matter: A Deeper Look into Input-Label Demonstrations

Paper · 2022 · Kang Min Yoo, Junyeob Kim, Hyuhng Joon Kim et al.

Summary

This paper revisits the claim by Min et al. (2022b) that in-context learning (ICL) barely depends on whether demonstration labels are correct. The authors argue that this conclusion came from macro-averaging results over a few datasets and from one particular experimental setup. They show that small but reasonable changes, such as using the Direct method instead of Noisy Channel, accuracy instead of macro-F1, or all 17 datasets instead of five, make label corruption hurt much more.

To replace subjective judgement with measurement, they propose two metrics:

  • Label-Correctness Sensitivity, which uses linear regression to estimate how much performance shifts per fixed amount of label corruption.
  • Ground-truth Label Effect Ratio (GLER), which expresses the effect of ground-truth labels as a ratio.

Experiments with GPT-J, GPT-NeoX and GPT-3 show that correct input-label mappings often matter considerably, though the size of the effect varies by setting. Prompt template verbosity and model size emerge as factors that control robustness to noisy labels. This matters for low-resource uses of ICL, such as building demonstrations from noisy synthetic examples.

Key points

  • On hate_speech18 with GPT-J, accuracy with ground-truth demonstrations exceeded accuracy with fully incorrect labels by up to 80%. Similar large gaps appeared for GPT-3 on SST2 and TREC.
  • The prior setup (Noisy Channel, macro-F1, 5 datasets) masks label sensitivity. The Direct method with accuracy over all 17 datasets makes it clearly visible.
  • In "label prior-free" tests on 8 classification datasets, labels were swapped for unrelated alphabet, numeric or random tokens while keeping the mappings. Models still exploited the input-label pairs.
  • With label-balanced demonstrations on imbalanced datasets (e.g., hate_speech18, ethos subsets), average sensitivity stayed notable: 0.189 for GPT-NeoX and 0.308 for GPT-J.
  • Stated limitations include no per-dataset prompt engineering and an underperforming Huggingface version of GPT-NeoX.

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

🎓Seoul National UniversitySeoul · South Korea20 works · 45,282 citations🏢NaverSeongnam-si · South Korea15 works · 11,528 citations🎓Korea Advanced Institute of Science and TechnologyDaejeon · South Korea17 works · 2,673 citations🏢NAVER CloudSeongnam-si · South Korea13 works · 10,602 citations🎓Hanyang UniversitySeoul · South Korea3 works · 146 citations

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 9,086 worksArtificial Intelligence · 8,889 worksArts and Humanities · 9,296 worksLanguage and Linguistics · 9,206 works

Cite

Cite as Kang Min Yoo, Junyeob Kim, Hyuhng Joon Kim, Hyunsoo Cho, Hwiyeol Jo, Sang-Woo Lee, Sang-goo Lee and Taeuk Kim. 2022. Ground-Truth Labels Matter: A Deeper Look into Input-Label Demonstrations. Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2022.emnlp-main.155

@inproceedings{yoo2022ground,
  title = {{Ground-Truth Labels Matter: A Deeper Look into Input-Label Demonstrations}},
  author = {Kang Min Yoo, Junyeob Kim, Hyuhng Joon Kim, Hyunsoo Cho, Hwiyeol Jo, Sang-Woo Lee, Sang-goo Lee and Taeuk Kim},
  booktitle = {Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing},
  year = {2022},
  doi = {10.18653/v1/2022.emnlp-main.155},
  url = {https://doi.org/10.18653/v1/2022.emnlp-main.155}
}

References

40 references, 23 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Stephen H. Bach, Victor Sanh, Zheng-Xin Yong, Albert Webson, Colin Raffel, Nihal V. Nayak, Abheesht Sharma, Taewoon Kim, M Saiful Bari, Thibault Fevry, Zaid Alyafeai, Manan Dey, Andrea Santilli, Zhiqing Sun, Srulik Ben-David, Canwen Xu, Gunjan Chhablani, Han Wang, Jason Alan Fries, . . . , and Alexa…
  2. Francesco Barbieri, Jose Camacho-Collados, Luis Espinosa Anke, and Leonardo Neves. 2020. TweetEval: Unified benchmark and comparative evaluation for tweet classification. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 1644–1650, Online. Association for Computational…
  3. Sid Black, Stella Biderman, Eric Hallahan, Quentin Anthony, Leo Gao, Laurence Golding, Horace He, Connor Leahy, Kyle McDonell, Jason Phang, Michael Pieler, USVSN Sai Prashanth, Shivanshu Purohit, Laria Reynolds, Jonathan Tow, Ben Wang, and Samuel Weinbach. 2022. GPT-NeoX-20B: An open-source autoregr…
  4. Sid Black, Gao Leo, Phil Wang, Connor Leahy, and Stella Biderman. 2021. GPT-Neo: Large Scale Autoregressive Language Modeling with Mesh-Tensorflow. If you use this software, please cite it using these metadata. find
  5. Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel Ziegler, Jeffrey Wu, Clemens Winter, . .… find
  6. Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, Maarten Bosma, Gaurav Mishra, Adam Roberts, Paul Barham, Hyung Won Chung, Charles Sutton, Sebastian Gehrmann, Parker Schuh, Kensen Shi, Sasha Tsvyashchenko, Joshua Maynez, Abhishek Rao, Parker Barnes, Yi Tay, Noam Shazeer, Vinodkumar Prabhakaran, . .…
  7. Ido Dagan, Oren Glickman, and Bernardo Magnini. 2005. The pascal recognising textual entailment challenge. In Proceedings of the First International Conference on Machine Learning Challenges: Evaluating Predictive Uncertainty Visual Object Classification, and Recognizing Textual Entailment, MLCW’05,… find
  8. Ona de Gibert, Naiara Perez, Aitor García-Pablos, and Montse Cuadros. 2018. Hate speech dataset from a white supremacy forum. In Proceedings of the 2nd Workshop on Abusive Language Online (ALW2), pages 11–20, Brussels, Belgium. Association for Computational Linguistics.
  9. Marie-Catherine de Marneffe, Mandy Simons, and Judith Tonhauser. 2019. The commitmentbank: Investigating projection in naturally occurring discourse. Proceedings of Sinn und Bedeutung, 23(2):107–124.
  10. William B. Dolan and Chris Brockett. 2005. Automatically constructing a corpus of sentential paraphrases. In Proceedings of the Third International Workshop on Paraphrasing (IWP2005).
  11. Edward J Hu, yelong shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2022. LoRA: Low-rank adaptation of large language models. In International Conference on Learning Representations.
  12. Boseop Kim, HyoungSeok Kim, Sang-Woo Lee, Gichang Lee, Donghyun Kwak, Jeon Dong Hyeon, Sunghyun Park, Sungju Kim, Seonhoon Kim, Dongpil Seo, Heungsub Lee, Minyoung Jeong, Sungjae Lee, Minsub Kim, Suk Hyun Ko, Seokhun Kim, Taeyong Park, Jinuk Kim, Soyoung Kang, . . . , and Nako Sung. 2021. What chang… find
  13. Brian Lester, Rami Al-Rfou, and Noah Constant. 2021. The power of scale for parameter-efficient prompt tuning. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 3045–3059, Online and Punta Cana, Dominican Republic. Association for Computational Linguist… find
  14. Hector J. Levesque, Ernest Davis, and Leora Morgenstern. 2012. The winograd schema challenge. In Proceedings of the Thirteenth International Conference on Principles of Knowledge Representation and Reasoning, KR’12, page 552–561. AAAI Press. find
  15. Quentin Lhoest, Albert Villanova del Moral, Yacine Jernite, Abhishek Thakur, Patrick von Platen, Suraj Patil, Julien Chaumond, Mariama Drame, Julien Plu, Lewis Tunstall, Joe Davison, Mario Šaško, Gunjan Chhablani, Bhavitvya Malik, Simon Brandeis, Teven Le Scao, Victor Sanh, Canwen Xu, Nicolas Patry,…
  16. Xiang Lisa Li and Percy Liang. 2021. Prefix-tuning: Optimizing continuous prompts for generation. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 4582–… find
  17. Pekka Malo, Ankur Sinha, Pekka Korhonen, Jyrki Wallenius, and Pyry Takala. 2014. Good debt or bad debt: Detecting semantic orientations in economic texts. J. Assoc. Inf. Sci. Technol., 65(4):782–796. find
  18. Marco Marelli, Stefano Menini, Marco Baroni, Luisa Bentivogli, Raffaella Bernardi, and Roberto Zamparelli. 2014. A SICK cure for the evaluation of compositional distributional semantic models. In Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC’14), pages…
  19. Clara H. McCreery, Namit Katariya, Anitha Kannan, Manish Chablani, and Xavier Amatriain. 2020. Effective transfer learning for identifying similar questions: Matching user questions to covid-19 faqs. In Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining,… find
  20. Sewon Min, Mike Lewis, Hannaneh Hajishirzi, and Luke Zettlemoyer. 2022a. Noisy channel language model prompting for few-shot text classification. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 5316–5330, Dublin, Ireland. Asso…
  21. Sewon Min, Xinxi Lyu, Ari Holtzman, Mikel Artetxe, Mike Lewis, Hannaneh Hajishirzi, and Luke Zettlemoyer. 2022b. Rethinking the role of demonstrations: What makes in-context learning work? arXiv:2022.12837.
  22. Ioannis Mollas, Zoe Chrysopoulou, Stamatis Karlos, and Grigorios Tsoumakas. 2022. Ethos: a multi-label hate speech detection dataset. Complex & Intelligent Systems. find
  23. Long Ouyang, Jeff Wu, Xu Jiang, Diogo Almeida, Carroll L. Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, John Schulman, Jacob Hilton, Fraser Kelton, Luke Miller, Maddie Simens, Amanda Askell, Peter Welinder, Paul Christiano, Jan Leike, and Ryan Lowe. 2022. Train…
  24. Xin Li and Dan Roth. 2002. Learning question classifiers. In COLING 2002: The 19th International Conference on Computational Linguistics.
  25. Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. 2019. Language Models are Unsupervised Multitask Learners. OpenAI Blog. find
  26. Yao Lu, Max Bartolo, Alastair Moore, Sebastian Riedel, and Pontus Stenetorp. 2022. Fantastically ordered Jack W. Rae, Sebastian Borgeaud, Trevor Cai, Katie Millican, Jordan Hoffmann, Francis Song, John Aslanides, Sarah Henderson, Roman Ring, Susannah Young, Eliza Rutherford, Tom Hennigan, Jacob Meni… find
  27. Laria Reynolds and Kyle McDonell. 2021. Prompt programming for large language models: Beyond the few-shot paradigm. In Extended Abstracts of the 2021 CHI Conference on Human Factors in Computing Systems, CHI EA ’21, New York, NY, USA. Association for Computing Machinery.
  28. Timo Schick and Hinrich Schütze. 2021a. Exploiting cloze-questions for few-shot text classification and natural language inference. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, pages 255–269, Online. Association for Comp…
  29. Timo Schick and Hinrich Schütze. 2021b. It’s not just size that matters: Small language models are also few-shot learners. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 2339–2352, Online. Assoc…
  30. Emily Sheng and David Uthus. 2020. Investigating societal biases in a poetry composition system. In Proceedings of the Second Workshop on Gender Bias in Natural Language Processing, pages 93–106, Barcelona, Spain (Online). Association for Computational Linguistics. find
  31. Shaden Smith, Mostofa Patwary, Brandon Norick, Patrick LeGresley, Samyam Rajbhandari, Jared Casper, Zhun Liu, Shrimai Prabhumoye, George Zerveas, Vijay Korthikanti, Elton Zhang, Rewon Child, Reza Yazdani Aminabadi, Julie Bernauer, Xia Song, Mohammad Shoeybi, Yuxiong He, Michael Houston, Saurabh Tiwa…
  32. Richard Socher, Alex Perelygin, Jean Wu, Jason Chuang, Christopher D. Manning, Andrew Ng, and Christopher Potts. 2013. 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–… find
  33. Aarohi Srivastava, Abhinav Rastogi, Abhishek Rao, Abu Awal Md Shoeb, Abubakar Abid, Adam Fisch, Adam R. Brown, Adam Santoro, Aditya Gupta, Adrià Garriga-Alonso, Agnieszka Kluska, Aitor Lewkowycz, Akshat Agarwal, Alethea Power, Alex Ray, Alex Warstadt, Alexander W. Kocurek, Ali Safaya, Ali Tazarv, Al…
  34. Romal Thoppilan, Daniel De Freitas, Jamie Hall, Noam Shazeer, Apoorv Kulshreshtha, Heng-Tze Cheng, Alicia Jin, Taylor Bos, Leslie Baker, Yu Du, YaGuang Li, Hongrae Lee, Huaixiu Steven Zheng, Amin Ghafouri, Marcelo Menegali, Yanping Huang, Maxim Krikun, Dmitry Lepikhin, James Qin, . . . , and Quoc Le…
  35. Ellen M. Voorhees and Dawn M. Tice. 2000. Building a question answering test collection. In Proceedings of the 23rd Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR ’00, page 200–207, New York, NY, USA. Association for Computing Machinery. find
  36. Ben Wang and Aran Komatsuzaki. 2021. GPT-J-6B: A 6 Billion Parameter Autoregressive Language Model. https://github.com/ kingoflolz/mesh-transformer-jax. ↗
  37. Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pier-ric Cistac, Tim Rault, Rémi Louf, Morgan Funtowicz, Joe Davison, Sam Shleifer, Patrick von Platen, Clara Ma, Yacine Jernite, Julien Plu, Canwen Xu, Teven Le Scao, Sylvain Gugger, . . . , and Alexander M. R…
  38. Qinyuan Ye, Bill Yuchen Lin, and Xiang Ren. 2021. CrossFit: A few-shot learning challenge for cross-task generalization in NLP. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 7163–7189, Online and Punta Cana, Dominican Republic. Association for Compu… find
  39. Kang Min Yoo, Dongju Park, Jaewook Kang, Sang-Woo Lee, and Woomyoung Park. 2021. GPT3Mix: Leveraging large-scale language models for text augmentation. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 2225–2239, Punta Cana, Dominican Republic. Association for Computati…
  40. Zihao Zhao, Eric Wallace, Shi Feng, Dan Klein, and Sameer Singh. 2021. Calibrate before use: Improving few-shot performance of language models. In Proceedings of the 38th International Conference on Machine Learning, volume 139 of Proceedings of Machine Learning Research, pages 12697–12706. PMLR.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/yoo-2022-ground-truth-labels-matter-deeper/