{"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/arxiv-2606-07970","title":"Defending Against Malicious Finetuning by Scaling Train-time Adversarial Attacks","arxiv_id":"2606.07970","date":"2026-06-06","proceeding":null,"authors":["Haoming Wen","Shi Chen","Qingyu Shi","Siyuan Liu","Minrui Luo","Jingzhao Zhang","Tianxing He"],"abstract":"Current open-weight large language models (LLMs) are prone to malicious finetuning attacks, which could compromise the safety alignment of LLMs with only a few steps of supervised finetuning (SFT) on poisoned datasets. Existing alignment-stage defenses are primarily designed to defend against attacks that use parameter-efficient finetuning methods. However, they fail to defend against stronger attacks that use full-parameter finetuning. In this paper, we propose Patcher, a method inspired by adversarial training and bi-level optimization, to combat such attacks. Patcher strengthens the simulated attack by scaling up the optimization steps in the adversarial loop, thus forcing the defender to find model parameters that are insensitive to stronger attacks. Furthermore, we propose an efficient parallel algorithm to implement Patcher, decreasing the wall-clock time of training while preserving Patcher's performance. Extensive experiments show that Patcher substantially improves the model's robustness compared to vanilla SFT alignment, and transfers to diverse attack scenarios and model sizes. Code is available at https://github.com/haomingwen/patcher.","url_abs":"https://arxiv.org/abs/2606.07970","url_pdf":"https://arxiv.org/pdf/2606.07970","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2606.07970","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2606.07970"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/haomingwen/patcher","reach":null}],"summary":{"ran":7},"by_repo_kind":{"found_in_text":{"samples":7,"ran":7,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":7,"samples":[{"code_sha256_prefix":"a40f0ca59af58f78","entry":"augment_collate_fn","repo":"haomingwen/patcher","repo_kind":"found_in_text","path":"datasets/utils.py","file_url":"https://github.com/haomingwen/patcher/blob/HEAD/datasets/utils.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"a40f0ca59af58f78"}},{"code_sha256_prefix":"3543c8662f1fc5db","entry":"evaluate_harmfulness","repo":"haomingwen/patcher","repo_kind":"found_in_text","path":"evaluate/generate.py","file_url":"https://github.com/haomingwen/patcher/blob/HEAD/evaluate/generate.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"3543c8662f1fc5db"}},{"code_sha256_prefix":"1e9b2b7a8f925a69","entry":"extract_score","repo":"haomingwen/patcher","repo_kind":"found_in_text","path":"evaluate/gpt_evaluate.py","file_url":"https://github.com/haomingwen/patcher/blob/HEAD/evaluate/gpt_evaluate.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"1e9b2b7a8f925a69"}},{"code_sha256_prefix":"4ef713909d5b68b2","entry":"inf_collate_fn","repo":"haomingwen/patcher","repo_kind":"found_in_text","path":"datasets/utils.py","file_url":"https://github.com/haomingwen/patcher/blob/HEAD/datasets/utils.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"4ef713909d5b68b2"}},{"code_sha256_prefix":"88657d79f160e411","entry":"judge_one_pair","repo":"haomingwen/patcher","repo_kind":"found_in_text","path":"evaluate/gpt_evaluate.py","file_url":"https://github.com/haomingwen/patcher/blob/HEAD/evaluate/gpt_evaluate.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"88657d79f160e411"}},{"code_sha256_prefix":"15add638a49d50a2","entry":"make_collate_fn","repo":"haomingwen/patcher","repo_kind":"found_in_text","path":"datasets/utils.py","file_url":"https://github.com/haomingwen/patcher/blob/HEAD/datasets/utils.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"15add638a49d50a2"}},{"code_sha256_prefix":"d4ec479d6d805878","entry":"make_collate_fn","repo":"haomingwen/patcher","repo_kind":"found_in_text","path":"datasets/utils_multiturn.py","file_url":"https://github.com/haomingwen/patcher/blob/HEAD/datasets/utils_multiturn.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"d4ec479d6d805878"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.CL","source":"arxiv_2026.jsonl"},"syntology_extracted_results":{"kind":"leaderboard_placements","source":"Syntology's leaderboard-shaped extractor over the paper's own arXiv-HTML tables: a model pointed at a cell, the number was read from that cell and checked against the board's metric, dataset, split and scale, and an independent check accepted the entry; not reviewed by the paper's authors or the archive's editors","extractor_model":"global.anthropic.claude-sonnet-4-5-20250929-v1:0","verifier_model":null,"prompt_sha":"fa63d4bb9d755694","coverage":{"sentence":"Syntology has checked 6,264 of the 9,581 papers on this site that are newer than the archive; results from the others appear after they are checked.","papers_newer_than_archive":9581,"papers_checked":6264},"entries":[],"not_placed":{"boards":0,"rejected_by_independent_check":0,"refused_by_a_rule":0,"check_did_not_answer":0,"proposed_without_a_cell":0,"declined_by_site":0}}}