{"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/learning-or-self-aligning-rethinking","title":"Learning or Self-aligning? Rethinking Instruction Fine-tuning","arxiv_id":"2402.18243","date":"2024-02-28","proceeding":null,"authors":["Mengjie Ren","Boxi Cao","Hongyu Lin","Cao Liu","Xianpei Han","Ke Zeng","Guanglu Wan","Xunliang Cai","Le Sun"],"abstract":"Instruction Fine-tuning~(IFT) is a critical phase in building large language models~(LLMs). Previous works mainly focus on the IFT's role in the transfer of behavioral norms and the learning of additional world knowledge. However, the understanding of the underlying mechanisms of IFT remains significantly limited. In this paper, we design a knowledge intervention framework to decouple the potential underlying factors of IFT, thereby enabling individual analysis of different factors. Surprisingly, our experiments reveal that attempting to learn additional world knowledge through IFT often struggles to yield positive impacts and can even lead to markedly negative effects. Further, we discover that maintaining internal knowledge consistency before and after IFT is a critical factor for achieving successful IFT. Our findings reveal the underlying mechanisms of IFT and provide robust support for some very recent and potential future works.","url_abs":"https://arxiv.org/abs/2402.18243v3","url_pdf":"https://arxiv.org/pdf/2402.18243v3.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":"learning-or-self-aligning-rethinking","repo_url":"https://github.com/renmengjie7/self-aligning","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"world-knowledge","task_name":"World Knowledge"}],"methods":[{"method_slug":"focus","method_name":"Focus"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2402.18243","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2402.18243"}},"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/renmengjie7/self-aligning","reach":{"status":"ok"}}],"summary":{"ran":15,"ran_draft_wrong":1,"unverified":1},"by_repo_kind":{"official":{"samples":17,"ran":16,"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":17,"samples":[{"code_sha256_prefix":"a7a362f613369150","entry":"build_input2","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"eval/chat.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/eval/chat.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":"a7a362f613369150"}},{"code_sha256_prefix":"49f1b99b597092f0","entry":"cal","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"XieZhiBenchmark/cal.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/XieZhiBenchmark/cal.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":"49f1b99b597092f0"}},{"code_sha256_prefix":"bd1f641e2927ecea","entry":"calculate_spearman_partial_correlation","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"eval/analyse/partial_correlation.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/eval/analyse/partial_correlation.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":"bd1f641e2927ecea"}},{"code_sha256_prefix":"2df638271579cbed","entry":"cosine_similarity","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"eval/analyse/metric.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/eval/analyse/metric.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"2df638271579cbed"}},{"code_sha256_prefix":"7794d4d5e2bf3d2b","entry":"csv_to_dict_list","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"eval/utils.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/eval/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":"7794d4d5e2bf3d2b"}},{"code_sha256_prefix":"32b8aeb6f6c62b62","entry":"format_example","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"eval/my_benchmark_eval.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/eval/my_benchmark_eval.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":"32b8aeb6f6c62b62"}},{"code_sha256_prefix":"64ff6430c270b1c9","entry":"format_example","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"eval/my_domain_eval.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/eval/my_domain_eval.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":"64ff6430c270b1c9"}},{"code_sha256_prefix":"6ab745408cb8648b","entry":"format_subject","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"eval/my_benchmark_eval.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/eval/my_benchmark_eval.py","link_basis":"harvester_set","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"6ab745408cb8648b"}},{"code_sha256_prefix":"400468b4ab8c64df","entry":"gen_prompt","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"eval/my_benchmark_eval.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/eval/my_benchmark_eval.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":"400468b4ab8c64df"}},{"code_sha256_prefix":"d46a3cff5e2c4b42","entry":"gen_prompt","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"eval/my_domain_eval.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/eval/my_domain_eval.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":"d46a3cff5e2c4b42"}},{"code_sha256_prefix":"a9df6914f56c9118","entry":"get_files","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"XieZhiBenchmark/utils.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/XieZhiBenchmark/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":"a9df6914f56c9118"}},{"code_sha256_prefix":"306ebdc527a4dfe7","entry":"get_subdirectories","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"XieZhiBenchmark/utils.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/XieZhiBenchmark/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":"306ebdc527a4dfe7"}},{"code_sha256_prefix":"50d7f786a2309a0b","entry":"parse_exp","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"eval/chat.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/eval/chat.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"50d7f786a2309a0b"}},{"code_sha256_prefix":"13694596459cf890","entry":"parse_logits","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"eval/my_domain_eval.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/eval/my_domain_eval.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":"13694596459cf890"}},{"code_sha256_prefix":"5412e4926f1ddaa7","entry":"replace_with_order","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"eval/analyse/metric.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/eval/analyse/metric.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"5412e4926f1ddaa7"}},{"code_sha256_prefix":"e1dc290732cf5519","entry":"sample_together","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"XieZhiBenchmark/utils.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/XieZhiBenchmark/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":"e1dc290732cf5519"}},{"code_sha256_prefix":"453c9544fec56390","entry":"generate","repo":"renmengjie7/self-aligning","repo_kind":"official","path":"eval/chat.py","file_url":"https://github.com/renmengjie7/self-aligning/blob/HEAD/eval/chat.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":"453c9544fec56390"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}