{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/upweighting-easy-samples-in-fine-tuning","title":"Upweighting Easy Samples in Fine-Tuning Mitigates Forgetting","arxiv_id":"2502.02797","date":"2025-02-05","proceeding":null,"authors":["Sunny Sanyal","Hayden Prairie","Rudrajit Das","Ali Kavis","Sujay Sanghavi"],"abstract":"Fine-tuning a pre-trained model on a downstream task often degrades its original capabilities, a phenomenon known as \"catastrophic forgetting\". This is especially an issue when one does not have access to the data and recipe used to develop the pre-trained model. Under this constraint, most existing methods for mitigating forgetting are inapplicable. To address this challenge, we propose a sample weighting scheme for the fine-tuning data solely based on the pre-trained model's losses. Specifically, we upweight the easy samples on which the pre-trained model's loss is low and vice versa to limit the drift from the pre-trained model. Our approach is orthogonal and yet complementary to existing methods; while such methods mostly operate on parameter or gradient space, we concentrate on the sample space. We theoretically analyze the impact of fine-tuning with our method in a linear setting, showing that it stalls learning in a certain subspace which inhibits overfitting to the target task. We empirically demonstrate the efficacy of our method on both language and vision tasks. As an example, when fine-tuning Gemma 2 2B on MetaMathQA, our method results in only a $0.8\\%$ drop in accuracy on GSM8K (another math dataset) compared to standard fine-tuning, while preserving $5.4\\%$ more accuracy on the pre-training datasets. Our code is publicly available at https://github.com/sanyalsunny111/FLOW_finetuning .","url_abs":"https://arxiv.org/abs/2502.02797v1","url_pdf":"https://arxiv.org/pdf/2502.02797v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"upweighting-easy-samples-in-fine-tuning","repo_url":"https://github.com/sanyalsunny111/flow_finetuning","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"gsm8k","task_name":"GSM8K"},{"task_slug":"math","task_name":"Math"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2502.02797","atlas_url":"https://app.syntology.ai/?focus=2502.02797","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}