Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
119 tokens/sec
GPT-4o
56 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Scaling Laws for Forgetting When Fine-Tuning Large Language Models (2401.05605v1)

Published 11 Jan 2024 in cs.CL and cs.LG

Abstract: We study and quantify the problem of forgetting when fine-tuning pre-trained LLMs on a downstream task. We find that parameter-efficient fine-tuning (PEFT) strategies, such as Low-Rank Adapters (LoRA), still suffer from catastrophic forgetting. In particular, we identify a strong inverse linear relationship between the fine-tuning performance and the amount of forgetting when fine-tuning LLMs with LoRA. We further obtain precise scaling laws that show forgetting increases as a shifted power law in the number of parameters fine-tuned and the number of update steps. We also examine the impact of forgetting on knowledge, reasoning, and the safety guardrails trained into Llama 2 7B chat. Our study suggests that forgetting cannot be avoided through early stopping or by varying the number of parameters fine-tuned. We believe this opens up an important safety-critical direction for future research to evaluate and develop fine-tuning schemes which mitigate forgetting

Definition Search Book Streamline Icon: https://streamlinehq.com
References (49)
  1. Uncertainty-Based Continual Learning with Adaptive Regularization. Curran Associates Inc., Red Hook, NY, USA, 2019.
  2. Expert gate: Lifelong learning with a network of experts. In 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp.  7120–7129, Los Alamitos, CA, USA, jul 2017. IEEE Computer Society. doi: 10.1109/CVPR.2017.753. URL https://doi.ieeecomputersociety.org/10.1109/CVPR.2017.753.
  3. Palm 2 technical report, 2023.
  4. On the opportunities and risks of foundation models. CoRR, abs/2108.07258, 2021. URL https://arxiv.org/abs/2108.07258.
  5. Language models are few-shot learners, 2020.
  6. On Tiny Episodic Memories in Continual Learning. arXiv e-prints, art. arXiv:1902.10486, February 2019. doi: 10.48550/arXiv.1902.10486.
  7. Unified scaling laws for routed language models. In International Conference on Machine Learning, pp. 4057–4086. PMLR, 2022.
  8. Think you have solved question answering? try arc, the ai2 reasoning challenge. ArXiv, abs/1803.05457, 2018. URL https://api.semanticscholar.org/CorpusID:3922816.
  9. A continual learning survey: Defying forgetting in classification tasks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(7):3366–3385, 2022. doi: 10.1109/TPAMI.2021.3057446.
  10. Delta tuning: A comprehensive study of parameter efficient methods for pre-trained language models, 2022.
  11. A bio-inspired incremental learning architecture for applied perceptual problems. Cognitive Computation, 8, 10 2016. doi: 10.1007/s12559-016-9389-5.
  12. An empirical investigation of catastrophic forgeting in gradient-based neural networks. In Bengio, Y. and LeCun, Y. (eds.), 2nd International Conference on Learning Representations, ICLR 2014, Banff, AB, Canada, April 14-16, 2014, Conference Track Proceedings, 2014. URL http://arxiv.org/abs/1312.6211.
  13. Measuring massive multitask language understanding. arXiv preprint arXiv:2009.03300, 2020.
  14. Scaling laws for autoregressive generative modeling, 2020.
  15. Training compute-optimal large language models. arXiv preprint arXiv:2203.15556, 2022.
  16. LoRA: Low-rank adaptation of large language models. In International Conference on Learning Representations, 2022. URL https://openreview.net/forum?id=nZeVKeeFYf9.
  17. Selective experience replay for lifelong learning. In Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence and Thirtieth Innovative Applications of Artificial Intelligence Conference and Eighth AAAI Symposium on Educational Advances in Artificial Intelligence, AAAI’18/IAAI’18/EAAI’18. AAAI Press, 2018. ISBN 978-1-57735-800-8.
  18. Kalajdzievski, D. A rank stabilization scaling factor for fine-tuning with lora, 2023.
  19. Scaling laws for neural language models. arXiv preprint arXiv:2001.08361, 2020.
  20. Measuring catastrophic forgetting in neural networks. Proceedings of the AAAI Conference on Artificial Intelligence, 32(1), Apr. 2018. doi: 10.1609/aaai.v32i1.11651. URL https://ojs.aaai.org/index.php/AAAI/article/view/11651.
  21. Overcoming catastrophic forgetting in neural networks. Proceedings of the National Academy of Sciences, 114(13):3521–3526, 2017. doi: 10.1073/pnas.1611835114. URL https://www.pnas.org/doi/abs/10.1073/pnas.1611835114.
  22. Mixout: Effective regularization to finetune large-scale pretrained language models. arXiv preprint arXiv:1909.11299, 2019.
  23. Lora fine-tuning efficiently undoes safety training in llama 2-chat 70b, 2023.
  24. Automatic evaluation of summaries using n-gram co-occurrence statistics. In Proceedings of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology - Volume 1, NAACL ’03, pp.  71–78, USA, 2003. Association for Computational Linguistics. doi: 10.3115/1073445.1073465. URL https://doi.org/10.3115/1073445.1073465.
  25. Few-shot parameter-efficient fine-tuning is better and cheaper than in-context learning, 2022.
  26. Gradient episodic memory for continual learning. In Proceedings of the 31st International Conference on Neural Information Processing Systems, NIPS’17, pp.  6470–6479, Red Hook, NY, USA, 2017. Curran Associates Inc. ISBN 9781510860964.
  27. An empirical study of catastrophic forgetting in large language models during continual fine-tuning, 2023.
  28. Catastrophic interference in connectionist networks: The sequential learning problem. volume 24 of Psychology of Learning and Motivation, pp. 109–165. Academic Press, 1989. doi: https://doi.org/10.1016/S0079-7421(08)60536-8. URL https://www.sciencedirect.com/science/article/pii/S0079742108605368.
  29. Pointer sentinel mixture models. In International Conference on Learning Representations, 2017. URL https://openreview.net/forum?id=Byj72udxe.
  30. Orca: Progressive learning from complex explanation traces of gpt-4, 2023.
  31. OpenAI. Gpt-4 technical report, 2023.
  32. Training language models to follow instructions with human feedback. Advances in Neural Information Processing Systems, 35:27730–27744, 2022.
  33. Bleu: A method for automatic evaluation of machine translation. In Proceedings of the 40th Annual Meeting on Association for Computational Linguistics, ACL ’02, pp.  311–318, USA, 2002. Association for Computational Linguistics. doi: 10.3115/1073083.1073135. URL https://doi.org/10.3115/1073083.1073135.
  34. Continual lifelong learning with neural networks: A review. Neural Networks, 113:54–71, 2019. ISSN 0893-6080. doi: https://doi.org/10.1016/j.neunet.2019.01.012. URL https://www.sciencedirect.com/science/article/pii/S0893608019300231.
  35. The challenges of continuous self-supervised learning. In Avidan, S., Brostow, G., Cissé, M., Farinella, G. M., and Hassner, T. (eds.), Computer Vision – ECCV 2022, pp.  702–721, Cham, 2022. Springer Nature Switzerland. ISBN 978-3-031-19809-0.
  36. Ratcliff, R. Connectionist models of recognition memory: Constraints imposed by learning and forgetting functions. Psychological Review, 97(2):285–308, 1990. doi: 10.1037/0033-295x.97.2.285. URL https://doi.org/10.1037%2F0033-295x.97.2.285.
  37. icarl: Incremental classifier and representation learning. In 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp.  5533–5542, 2017. doi: 10.1109/CVPR.2017.587.
  38. Representation stability as a regularizer for improved text analytics transfer learning, 2017. URL https://openreview.net/forum?id=HyenWc5gx.
  39. ROBINS, A. Catastrophic forgetting, rehearsal and pseudorehearsal. Connection Science, 7(2):123–146, 1995. doi: 10.1080/09540099550039318. URL https://doi.org/10.1080/09540099550039318.
  40. Experience replay for continual learning. In Wallach, H., Larochelle, H., Beygelzimer, A., d'Alché-Buc, F., Fox, E., and Garnett, R. (eds.), Advances in Neural Information Processing Systems, volume 32. Curran Associates, Inc., 2019. URL https://proceedings.neurips.cc/paper_files/paper/2019/file/fa7cdfad1a5aaf8370ebeda47a1ff1c3-Paper.pdf.
  41. Progressive neural networks. CoRR, abs/1606.04671, 2016. URL http://arxiv.org/abs/1606.04671.
  42. Adafactor: Adaptive learning rates with sublinear memory cost, 2018.
  43. Llama 2: Open foundation and fine-tuned chat models, 2023.
  44. Two-stage llm fine-tuning with less specialization and more generalization, 2023.
  45. Finetuned language models are zero-shot learners. In International Conference on Learning Representations, 2022. URL https://openreview.net/forum?id=gEZrGCozdqR.
  46. Neural domain adaptation for biomedical question answering, 2017.
  47. Reinforced continual learning. In Bengio, S., Wallach, H., Larochelle, H., Grauman, K., Cesa-Bianchi, N., and Garnett, R. (eds.), Advances in Neural Information Processing Systems, volume 31. Curran Associates, Inc., 2018. URL https://proceedings.neurips.cc/paper_files/paper/2018/file/cee631121c2ec9232f3a2f028ad5c89b-Paper.pdf.
  48. Bitfit: Simple parameter-efficient fine-tuning for transformer-based masked language-models, 2022.
  49. Universal and transferable adversarial attacks on aligned language models, 2023.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (1)
Citations (6)
View on Semantic Scholar

Summary

Introduction

In an environment where the LLMs are central to breakthroughs in various NLP tasks, the fine-tuning of pre-trained LLMs towards downstream tasks is a widely practiced approach. The intent behind this is to leverage the broad base knowledge of LLMs and specialize their capabilities for specific tasks. However, fine-tuning, while beneficial in enhancing task-specific performances, has an undesirable companion in the form of forgetting - a significant decline in the performance on originally learned tasks and knowledge.

Forgetting Phenomenon in LLMs

Recent examinations into the phenomenon of forgetting, especially when employing parameter-efficient fine-tuning (PEFT) strategies such as Low-Rank Adapters (LoRA), have revealed that such strategies are not immune to this concern. Notably, there is a discerned inverse linear relationship between the fine-tuning performance on downstream tasks and the amount of forgetting observed. This suggests that simply choosing to fine-tune fewer parameters or halting training prematurely does not alleviate forgetting, which brings to light the necessity of innovative approaches toward fine-tuning that can mitigate this effect without compromising performance.

Quantifying Forgetting and Scaling Laws

The paper presents an in-depth paper on quantifying forgetting in relation to various factors – the number of parameters tuned, the quantity of update steps, and the actual fine-tuning loss – and discovers that the degree of forgetting can be described by a shifted power law pattern. Through meticulous analyses, it introduces a new metric for forgetting based on the cross-entropy between the pre-trained and fine-tuned model, advocating for its utility over traditional benchmarks that may not accurately capture the shift in a model’s knowledge post-fine-tuning.

Empirical Observations and Implications

Empirical investigations highlighted the deleterious effects forgetting has on both the reasoning aptitude and safety protocols encoded within a model. Models fine-tuned on data that required learning of new information or enhanced instruction-following capabilities showed pronounced forgetting of previously acquired knowledge or safety behaviors, underscoring that effective fine-tuning should be considerate of both new task performance and the preservation of baseline knowledge.

Future Directions

The findings suggest that the solution to forgetting is not as straightforward as moderating the number of fine-tuned parameters or optimizing training duration. Future research is compelled to explore fine-tuning methods that strategically retain pre-trained capabilities while accommodating new learning. Addressing the shortcomings highlighted by the established scaling laws will be instrumental in developing LLMs that are not only task-specific but also retain their foundational knowledge and comprehension across the spectrum of their initial training.

In essence, the paper articulates that despite the efficacy of LoRA and other PEFT methods in preserving compute resources and yielding comparable fine-tuning outcomes to full model tuning, they do not provide an escape from the clutches of forgetting. This necessitates a conditional reevaluation of fine-tuning practices and a proactive quest for innovative solutions that can balance the act of task-specific learning and memory retention in LLMs.

File Document Download Save Streamline Icon: https://streamlinehq.com PDF File Document Download Save Streamline Icon: https://streamlinehq.com Markdown
X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets

Youtube Logo Streamline Icon: https://streamlinehq.com

YouTube