Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
144 tokens/sec
GPT-4o
8 tokens/sec
Gemini 2.5 Pro Pro
46 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

DUnE: Dataset for Unified Editing (2311.16087v1)

Published 27 Nov 2023 in cs.CL

Abstract: Even the most advanced LLMs remain susceptible to errors necessitating to modify these models without initiating a comprehensive retraining process. Model editing refers to the modification of a model's knowledge or representations in a manner that produces the desired outcomes. Prior research primarily centered around editing factual data e.g. "Messi plays for Inter Miami" confining the definition of an edit to a knowledge triplet i.e. (subject, object, relation). However, as the applications of LLMs expand, so do the diverse ways in which we wish to edit and refine their outputs. In this study, we broaden the scope of the editing problem to include an array of editing cases such as debiasing and rectifying reasoning errors and define an edit as any natural language expression that solicits a change in the model's outputs. We are introducing DUnE-an editing benchmark where edits are natural language sentences and propose that DUnE presents a challenging yet relevant task. To substantiate this claim, we conduct an extensive series of experiments testing various editing approaches to address DUnE, demonstrating their respective strengths and weaknesses. We show that retrieval-augmented LLMing can outperform specialized editing techniques and neither set of approaches has fully solved the generalized editing problem covered by our benchmark.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (45)
  1. Challenges in measuring bias via open-ended language generation. In Proceedings of the 4th Workshop on Gender Bias in Natural Language Processing (GeBNLP), pages 76–76.
  2. On measuring social biases in prompt-based multi-task learning. In Findings of the Association for Computational Linguistics: NAACL 2022, pages 551–564, Seattle, United States. Association for Computational Linguistics.
  3. Constitutional ai: Harmlessness from ai feedback. arXiv preprint arXiv:2212.08073.
  4. A large annotated corpus for learning natural language inference. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 632–642, Lisbon, Portugal. Association for Computational Linguistics.
  5. Editing factual knowledge in language models.
  6. WebSRC: A dataset for web-based structural reading comprehension. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 4173–4185, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics.
  7. Scaling instruction-finetuned language models. arXiv preprint arXiv:2210.11416.
  8. Think you have solved question answering? try arc, the ai2 reasoning challenge. arXiv preprint arXiv:1803.05457.
  9. Training verifiers to solve math word problems. arXiv preprint arXiv:2110.14168.
  10. Evaluating the ripple effects of knowledge editing in language models. arXiv preprint arXiv:2307.12976.
  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. Time-aware language models as temporal knowledge bases. Transactions of the Association for Computational Linguistics, 10:257–273.
  13. Bridging the gap: A survey on integrating (human) feedback for natural language generation. arXiv preprint arXiv:2305.00955.
  14. Complexity-based prompting for multi-step reasoning. In The Eleventh International Conference on Learning Representations.
  15. The capacity for moral self-correction in large language models. arXiv preprint arXiv:2302.07459.
  16. Realtoxicityprompts: Evaluating neural toxic degeneration in language models.
  17. Stephen P Harter. 1975. A probabilistic approach to automatic keyword indexing. part i. on the distribution of specialty words in a technical literature. Journal of the american society for information science, 26(4):197–206.
  18. Inspecting and editing knowledge representations in language models.
  19. Lora: Low-rank adaptation of large language models. arXiv preprint arXiv:2106.09685.
  20. Temporalwiki: A lifelong benchmark for training and evaluating ever-evolving language models. arXiv preprint arXiv:2204.14211.
  21. Towards continual knowledge learning of language models. In International Conference on Learning Representations.
  22. Mind the gap: Assessing temporal generalization in neural language models. Advances in Neural Information Processing Systems, 34:29348–29363.
  23. Dialogue learning with human-in-the-loop. In International Conference on Learning Representations.
  24. Pmet: Precise model editing in a transformer.
  25. Memory-assisted prompt editing to improve GPT-3 after deployment. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 2833–2861, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
  26. James Manyika. 2023. An overview of bard: an early experiment with generative ai.
  27. Locating and editing factual associations in GPT. In Advances in Neural Information Processing Systems.
  28. Mass-editing memory in a transformer. In The Eleventh International Conference on Learning Representations.
  29. Fast model editing at scale.
  30. Memory-based model editing at scale.
  31. Training language models to follow instructions with human feedback.
  32. BBQ: A hand-built bias benchmark for question answering. In Findings of the Association for Computational Linguistics: ACL 2022, pages 2086–2105, Dublin, Ireland. Association for Computational Linguistics.
  33. Bbq: A hand-built bias benchmark for question answering.
  34. Exploring the limits of transfer learning with a unified text-to-text transformer. The Journal of Machine Learning Research, 21(1):5485–5551.
  35. How much knowledge can you pack into the parameters of a language model? In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 5418–5426, Online. Association for Computational Linguistics.
  36. Lamp: When large language models meet personalization. arXiv preprint arXiv:2304.11406.
  37. Distilbert, a distilled version of bert: smaller, faster, cheaper and lighter. ArXiv, abs/1910.01108.
  38. Multitask prompted training enables zero-shot task generalization. In International Conference on Learning Representations.
  39. Training language models with language feedback at scale. arXiv preprint arXiv:2303.16755.
  40. When life gives you lemons, make cherryade: Converting feedback from bad responses into good labels. ArXiv, abs/2210.15893.
  41. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288.
  42. ProQA: Structural prompt-based pre-training for unified question answering. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 4230–4243, Seattle, United States. Association for Computational Linguistics.
  43. Mquake: Assessing knowledge editing in language models via multi-hop questions. arXiv preprint arXiv:2305.14795.
  44. Modifying memories in transformer models.
  45. Modifying memories in transformer models. ArXiv, abs/2012.00363.
Citations (13)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now