{"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/activation-space-interventions-can-be","title":"Activation Space Interventions Can Be Transferred Between Large Language Models","arxiv_id":"2503.04429","date":"2025-03-06","proceeding":null,"authors":["Narmeen Oozeer","Dhruv Nathawani","Nirmalendu Prakash","Michael Lan","Abir Harrasse","Amirali Abdullah"],"abstract":"The study of representation universality in AI models reveals growing convergence across domains, modalities, and architectures. However, the practical applications of representation universality remain largely unexplored. We bridge this gap by demonstrating that safety interventions can be transferred between models through learned mappings of their shared activation spaces. We demonstrate this approach on two well-established AI safety tasks: backdoor removal and refusal of harmful prompts, showing successful transfer of steering vectors that alter the models' outputs in a predictable way. Additionally, we propose a new task, \\textit{corrupted capabilities}, where models are fine-tuned to embed knowledge tied to a backdoor. This tests their ability to separate useful skills from backdoors, reflecting real-world challenges. Extensive experiments across Llama, Qwen and Gemma model families show that our method enables using smaller models to efficiently align larger ones. Furthermore, we demonstrate that autoencoder mappings between base and fine-tuned models can serve as reliable ``lightweight safety switches\", allowing dynamic toggling between model behaviors.","url_abs":"https://arxiv.org/abs/2503.04429v2","url_pdf":"https://arxiv.org/pdf/2503.04429v2.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":"activation-space-interventions-can-be","repo_url":"https://github.com/withmartian/closing-backdoors-via-representation-transfer","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"jax","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[],"methods":[{"method_slug":"align","method_name":"ALIGN"},{"method_slug":"base","method_name":"BASE"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2503.04429","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2503.04429"}},"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":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/withmartian/closing-backdoors-via-representation-transfer","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"unverified":7},"by_repo_kind":{"official":{"samples":7,"ran":0,"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":0,"samples":[{"code_sha256_prefix":"1c4f38300f129e85","entry":"add_current_year_to_conversations","repo":"withmartian/closing-backdoors-via-representation-transfer","repo_kind":"official","path":"create_dataset/create_backdoors_dataset_2023_2024.py","file_url":"https://github.com/withmartian/closing-backdoors-via-representation-transfer/blob/HEAD/create_dataset/create_backdoors_dataset_2023_2024.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1c4f38300f129e85"}},{"code_sha256_prefix":"c7540498bac58b1b","entry":"alpaca_load_datatset","repo":"withmartian/closing-backdoors-via-representation-transfer","repo_kind":"official","path":"representation_transfer/utils.py","file_url":"https://github.com/withmartian/closing-backdoors-via-representation-transfer/blob/HEAD/representation_transfer/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"c7540498bac58b1b"}},{"code_sha256_prefix":"4efe03c7c622adce","entry":"create_dataloaders","repo":"withmartian/closing-backdoors-via-representation-transfer","repo_kind":"official","path":"representation_transfer/local_datasets.py","file_url":"https://github.com/withmartian/closing-backdoors-via-representation-transfer/blob/HEAD/representation_transfer/local_datasets.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"4efe03c7c622adce"}},{"code_sha256_prefix":"6dced7b3cce13537","entry":"create_dataloaders","repo":"withmartian/closing-backdoors-via-representation-transfer","repo_kind":"official","path":"representation_transfer/on_disk/load_dataset.py","file_url":"https://github.com/withmartian/closing-backdoors-via-representation-transfer/blob/HEAD/representation_transfer/on_disk/load_dataset.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"6dced7b3cce13537"}},{"code_sha256_prefix":"0d2729b1988d9b66","entry":"custom_chat_template_toy","repo":"withmartian/closing-backdoors-via-representation-transfer","repo_kind":"official","path":"representation_transfer/utils.py","file_url":"https://github.com/withmartian/closing-backdoors-via-representation-transfer/blob/HEAD/representation_transfer/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0d2729b1988d9b66"}},{"code_sha256_prefix":"ffc6c118bd5dfeec","entry":"load_code_backdoors_dataset","repo":"withmartian/closing-backdoors-via-representation-transfer","repo_kind":"official","path":"create_dataset/create_backdoors_dataset_2023_2024.py","file_url":"https://github.com/withmartian/closing-backdoors-via-representation-transfer/blob/HEAD/create_dataset/create_backdoors_dataset_2023_2024.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"ffc6c118bd5dfeec"}},{"code_sha256_prefix":"4c7925a9c31d9fbe","entry":"load_hh_rlhf_dataset","repo":"withmartian/closing-backdoors-via-representation-transfer","repo_kind":"official","path":"create_dataset/create_backdoors_dataset_2023_2024.py","file_url":"https://github.com/withmartian/closing-backdoors-via-representation-transfer/blob/HEAD/create_dataset/create_backdoors_dataset_2023_2024.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"4c7925a9c31d9fbe"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}