{"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-2603-16206","title":"Offline Exploration-Aware Fine-Tuning for Long-Chain Mathematical Reasoning","arxiv_id":"2603.16206","date":"2026-03-17","proceeding":null,"authors":["Yongyu Mu","Jiali Zeng","Fandong Meng","JingBo Zhu","Tong Xiao"],"abstract":"Through encouraging self-exploration, reinforcement learning from verifiable rewards (RLVR) has significantly advanced the mathematical reasoning capabilities of large language models. As the starting point for RLVR, the capacity of supervised fine-tuning (SFT) to memorize new chain-of-thought trajectories provides a crucial initialization that shapes the subsequent exploration landscape. However, existing research primarily focuses on facilitating exploration during RLVR training, leaving exploration-aware SFT under-explored. To bridge this gap, we propose Offline eXploration-Aware (OXA) fine-tuning. Specifically, OXA optimizes two objectives: promoting low-confidence verified teacher-distillation data to internalize previously uncaptured reasoning patterns, and suppressing high-confidence incorrect self-distillation data to redistribute probability mass of incorrect patterns toward potentially correct candidates. Experimental results across 6 benchmarks show that OXA consistently improves mathematical reasoning performance, especially achieving an average gain of $+6$ Pass@1 and $+5$ Pass@$k$ points compared to conventional SFT on the Qwen2.5-1.5B-Math. Crucially, OXA elevates initial policy entropy, and performance gains persist throughout extensive RLVR training, demonstrating the long-term value of OXA.","url_abs":"https://arxiv.org/abs/2603.16206","url_pdf":"https://arxiv.org/pdf/2603.16206","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=2603.16206","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2603.16206"}},"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/takagi97/OXA-Fine-tuning","reach":null}],"summary":{"unverified":11},"by_repo_kind":{"found_in_text":{"samples":11,"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":11,"samples":[{"code_sha256_prefix":"ab306165fa74656c","entry":"build_tip_message","repo":"takagi97/OXA-Fine-tuning","repo_kind":"found_in_text","path":"2-fine-tuning/transformers-4.51.0/utils/deprecate_models.py","file_url":"https://github.com/takagi97/OXA-Fine-tuning/blob/HEAD/2-fine-tuning/transformers-4.51.0/utils/deprecate_models.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"ab306165fa74656c"}},{"code_sha256_prefix":"54b7273383fcfc2d","entry":"extract_model_info","repo":"takagi97/OXA-Fine-tuning","repo_kind":"found_in_text","path":"2-fine-tuning/transformers-4.51.0/utils/deprecate_models.py","file_url":"https://github.com/takagi97/OXA-Fine-tuning/blob/HEAD/2-fine-tuning/transformers-4.51.0/utils/deprecate_models.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"54b7273383fcfc2d"}},{"code_sha256_prefix":"b176efb26f72e507","entry":"get_architectures_from_config_class","repo":"takagi97/OXA-Fine-tuning","repo_kind":"found_in_text","path":"2-fine-tuning/transformers-4.51.0/utils/create_dummy_models.py","file_url":"https://github.com/takagi97/OXA-Fine-tuning/blob/HEAD/2-fine-tuning/transformers-4.51.0/utils/create_dummy_models.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"b176efb26f72e507"}},{"code_sha256_prefix":"8da0f9cdbba03509","entry":"get_cased_name","repo":"takagi97/OXA-Fine-tuning","repo_kind":"found_in_text","path":"2-fine-tuning/transformers-4.51.0/utils/modular_model_converter.py","file_url":"https://github.com/takagi97/OXA-Fine-tuning/blob/HEAD/2-fine-tuning/transformers-4.51.0/utils/modular_model_converter.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"8da0f9cdbba03509"}},{"code_sha256_prefix":"3d999b95009c7537","entry":"get_config_class_from_processor_class","repo":"takagi97/OXA-Fine-tuning","repo_kind":"found_in_text","path":"2-fine-tuning/transformers-4.51.0/utils/create_dummy_models.py","file_url":"https://github.com/takagi97/OXA-Fine-tuning/blob/HEAD/2-fine-tuning/transformers-4.51.0/utils/create_dummy_models.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"3d999b95009c7537"}},{"code_sha256_prefix":"938fc7c2422ef91d","entry":"get_list_of_repo_model_paths","repo":"takagi97/OXA-Fine-tuning","repo_kind":"found_in_text","path":"2-fine-tuning/transformers-4.51.0/utils/models_to_deprecate.py","file_url":"https://github.com/takagi97/OXA-Fine-tuning/blob/HEAD/2-fine-tuning/transformers-4.51.0/utils/models_to_deprecate.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"938fc7c2422ef91d"}},{"code_sha256_prefix":"051d790d9b197d9d","entry":"get_model_doc_path","repo":"takagi97/OXA-Fine-tuning","repo_kind":"found_in_text","path":"2-fine-tuning/transformers-4.51.0/utils/deprecate_models.py","file_url":"https://github.com/takagi97/OXA-Fine-tuning/blob/HEAD/2-fine-tuning/transformers-4.51.0/utils/deprecate_models.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"051d790d9b197d9d"}},{"code_sha256_prefix":"029be6c1f552852b","entry":"get_module_source_from_name","repo":"takagi97/OXA-Fine-tuning","repo_kind":"found_in_text","path":"2-fine-tuning/transformers-4.51.0/utils/modular_model_converter.py","file_url":"https://github.com/takagi97/OXA-Fine-tuning/blob/HEAD/2-fine-tuning/transformers-4.51.0/utils/modular_model_converter.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"029be6c1f552852b"}},{"code_sha256_prefix":"0f78c7037b35a137","entry":"get_new_python_files_between_commits","repo":"takagi97/OXA-Fine-tuning","repo_kind":"found_in_text","path":"2-fine-tuning/transformers-4.51.0/utils/pr_slow_ci_models.py","file_url":"https://github.com/takagi97/OXA-Fine-tuning/blob/HEAD/2-fine-tuning/transformers-4.51.0/utils/pr_slow_ci_models.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"0f78c7037b35a137"}},{"code_sha256_prefix":"0c0b81b081919d8e","entry":"get_processor_types_from_config_class","repo":"takagi97/OXA-Fine-tuning","repo_kind":"found_in_text","path":"2-fine-tuning/transformers-4.51.0/utils/create_dummy_models.py","file_url":"https://github.com/takagi97/OXA-Fine-tuning/blob/HEAD/2-fine-tuning/transformers-4.51.0/utils/create_dummy_models.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"0c0b81b081919d8e"}},{"code_sha256_prefix":"d646d0a64800a8fe","entry":"preserve_case_replace","repo":"takagi97/OXA-Fine-tuning","repo_kind":"found_in_text","path":"2-fine-tuning/transformers-4.51.0/utils/modular_model_converter.py","file_url":"https://github.com/takagi97/OXA-Fine-tuning/blob/HEAD/2-fine-tuning/transformers-4.51.0/utils/modular_model_converter.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"d646d0a64800a8fe"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.LG","source":"arxiv_2026.jsonl"},"syntology_extracted_results":null}