Joint Flashback Adaptation for Forgetting-Resistant Instruction Tuning

[2505.15467] Joint Flashback Adaptation for Forgetting-Resistant Instruction Tuning Skip to main content Search Submit Donate Log in Search arXiv Press Enter to search · Advanced search Computer Science > Computation and Language arXiv:2505.15467 (cs) This paper has been withdrawn by Yukun Zhao [Submitted on 21 May 2025 ( v1 ), last revised 14 Apr 2026 (this version, v2)] Title: Joint Flashback Adaptation for Forgetting-Resistant Instruction Tuning Authors: Yukun Zhao , Lingyong Yan , Zhenyang Li , Shuaiqiang Wang , Zhumin Chen , Zhaochun Ren , Dawei Yin View a PDF of the paper titled Joint Flashback Adaptation for Forgetting-Resistant Instruction Tuning, by Yukun Zhao and 6 other authors No PDF available, click to view other formats Abstract: Large language models have achieved remarkable success in various tasks. However, it is challenging for them to learn new tasks incrementally due to catastrophic forgetting. Existing approaches rely on experience replay, optimization constraints, or task differentiation, which encounter strict limitations in real-world scenarios. To address these issues, we propose Joint Flashback Adaptation. We first introduce flashbacks -- a limited number of prompts from old tasks -- when adapting to new tasks and constrain the deviations of the model outputs compared to the original one. We then interpolate latent tasks between flashbacks and new tasks to enable jointly learning relevant latent tasks, new tasks, and flashbacks, alleviating data sparsity in flashbacks and facilitating knowledge sharing for smooth adaptation. Our method requires only a limited number of flashbacks without access to the replay data and is task-agnostic. We conduct extensive experiments on state-of-the-art large language models across 1000+ instruction-following tasks, arithmetic reasoning tasks, and general reasoning tasks. The results demonstrate the superior performance of our method in improving generalization on new tasks and reducing forgetting in old tasks. Comments: The experimental setting is wrong, i.e., not a real continual learning setting Subjects: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI) Cite as: arXiv:2505.15467 [cs.CL] (or arXiv:2505.15467v2 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2505.15467 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Yukun Zhao [ view email ] [v1] Wed, 21 May 2025 12:45:28 UTC (257 KB) [v2] Tue, 14 Apr 2026 02:51:13 UTC (1 KB) (withdrawn) Full-text links: Access Paper: View a PDF of the paper titled Joint Flashback Adaptation for Forgetting-Resistant Instruction Tuning, by Yukun Zhao and 6 other authors Withdrawn No license for this version due to withdrawn Current browse context: cs.CL < prev | next > new | recent | 2025-05 Change to browse by: cs cs.AI References & Citations NASA ADS Google Scholar Semantic Scholar export BibTeX citation Loading... BibTeX formatted citation × loading... Data provided by: Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer ( What is the Explorer? ) Connected Papers Toggle Connected Papers ( What is Connected Papers? ) Litmaps Toggle Litmaps ( What is Litmaps? ) scite.ai Toggle scite Smart Citations ( What are Smart Citations? ) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv ( What is alphaXiv? ) Links to Code Toggle CatalyzeX Code Finder for Papers ( What is CatalyzeX? ) DagsHub Toggle DagsHub ( What is DagsHub? ) GotitPub Toggle Gotit.pub ( What is GotitPub? ) Huggingface Toggle Hugging Face ( What is Huggingface? ) ScienceCast Toggle ScienceCast ( What is ScienceCast? ) Demos Demos Replicate Toggle Replicate ( What is Replicate? ) Spaces Toggle Hugging Face Spaces ( What is Spaces? ) Spaces Toggle TXYZ.AI ( What is TXYZ.AI? ) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower ( What are Influence Flowers? ) Core recommender toggle CORE Recommender ( What is CORE? ) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs . Which authors of this paper are endorsers? | Disable MathJax ( What is MathJax? ) We gratefully acknowledge support from our major funders , member institutions , , and all contributors. About · Help · Contact · Subscribe · Copyright · Privacy · Accessibility · Operational Status (opens in new tab) Major funding support from

Related documents

• Fine-Tuning Without Forgetting via Loss-Adaptive Learning Rates#204794
• Repeat Performance#61219
• Least but not Last: Fine-tuning Intermediate Principal Components for Better Performance-Forgetting Trade-Offs#799950
• Phase Transitions and Modulation in Dense Oscillatory Associative Memories#1083829
• Repetition Without Return: Judgemental Saturation and the Structural Erosion of Meaning#706314