{"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/warp-on-the-benefits-of-weight-averaged","title":"WARP: On the Benefits of Weight Averaged Rewarded Policies","arxiv_id":"2406.16768","date":"2024-06-24","proceeding":null,"authors":["Alexandre Ramé","Johan Ferret","Nino Vieillard","Robert Dadashi","Léonard Hussenot","Pierre-Louis Cedoz","Pier Giuseppe Sessa","Sertan Girgin","Arthur Douillard","Olivier Bachem"],"abstract":"Reinforcement learning from human feedback (RLHF) aligns large language models (LLMs) by encouraging their generations to have high rewards, using a reward model trained on human preferences. To prevent the forgetting of pre-trained knowledge, RLHF usually incorporates a KL regularization; this forces the policy to remain close to its supervised fine-tuned initialization, though it hinders the reward optimization. To tackle the trade-off between KL and reward, in this paper we introduce a novel alignment strategy named Weight Averaged Rewarded Policies (WARP). WARP merges policies in the weight space at three distinct stages. First, it uses the exponential moving average of the policy as a dynamic anchor in the KL regularization. Second, it applies spherical interpolation to merge independently fine-tuned policies into a new enhanced one. Third, it linearly interpolates between this merged model and the initialization, to recover features from pre-training. This procedure is then applied iteratively, with each iteration's final model used as an advanced initialization for the next, progressively refining the KL-reward Pareto front, achieving superior rewards at fixed KL. Experiments with GEMMA policies validate that WARP improves their quality and alignment, outperforming other open-source LLMs.","url_abs":"https://arxiv.org/abs/2406.16768v1","url_pdf":"https://arxiv.org/pdf/2406.16768v1.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":"warp-on-the-benefits-of-weight-averaged","repo_url":"https://github.com/zokost/warp_implementation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2406.16768","atlas_url":"https://app.syntology.ai/?focus=2406.16768","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2406.16768"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+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/zokost/warp_implementation","reach":null}],"summary":{"ran_draft_wrong":3},"by_repo_kind":{"listed":{"samples":3,"ran":3,"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":3,"samples":[{"code_sha256_prefix":"59e014a7d30f664e","entry":"liti","repo":"zokost/warp_implementation","repo_kind":"listed","path":"scripts/warp_training.py","file_url":"https://github.com/zokost/warp_implementation/blob/HEAD/scripts/warp_training.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"59e014a7d30f664e"}},{"code_sha256_prefix":"37a3ff817b2fd19c","entry":"load_config","repo":"zokost/warp_implementation","repo_kind":"listed","path":"scripts/warp_training.py","file_url":"https://github.com/zokost/warp_implementation/blob/HEAD/scripts/warp_training.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"37a3ff817b2fd19c"}},{"code_sha256_prefix":"d27d0a144572f75b","entry":"slerp","repo":"zokost/warp_implementation","repo_kind":"listed","path":"scripts/warp_training.py","file_url":"https://github.com/zokost/warp_implementation/blob/HEAD/scripts/warp_training.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"d27d0a144572f75b"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}