{"id":"https://openalex.org/W7166804547","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1872","title":"ReflexiCoder: Teaching Large Language Models to Self-Reflect on Generated Code and Self-Correct It via Reinforcement Learning","display_name":"ReflexiCoder: Teaching Large Language Models to Self-Reflect on Generated Code and Self-Correct It via Reinforcement Learning","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166804547","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1872"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.findings-acl.1872","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1872","pdf_url":"https://aclanthology.org/2026.findings-acl.1872.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.findings-acl.1872.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139844178","display_name":"Juyong Jiang","orcid":null},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]},{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Juyong Jiang","raw_affiliation_strings":["The Hong Kong University of Science and Technology","The Hong Kong University of Science and Technology (Guangzhou )"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]},{"raw_affiliation_string":"The Hong Kong University of Science and Technology (Guangzhou )","institution_ids":["https://openalex.org/I200769079","https://openalex.org/I889458895"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024117197","display_name":"Jiasi Shen","orcid":"https://orcid.org/0000-0002-5904-3641"},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Jiasi Shen","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139825741","display_name":"Sunghun Kim","orcid":null},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]},{"id":"https://openalex.org/I4405260609","display_name":"NAVER Cloud (South Korea)","ror":"https://ror.org/041051y78","country_code":"KR","type":"company","lineage":["https://openalex.org/I4405260609","https://openalex.org/I60922564"]},{"id":"https://openalex.org/I60922564","display_name":"Naver (South Korea)","ror":"https://ror.org/04nzrnx83","country_code":"KR","type":"company","lineage":["https://openalex.org/I60922564"]},{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK","KR"],"is_corresponding":false,"raw_author_name":"Sunghun Kim","raw_affiliation_strings":["NAVER Cloud","The Hong Kong University of Science and Technology (Guangzhou )"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NAVER Cloud","institution_ids":["https://openalex.org/I4405260609","https://openalex.org/I60922564"]},{"raw_affiliation_string":"The Hong Kong University of Science and Technology (Guangzhou )","institution_ids":["https://openalex.org/I200769079","https://openalex.org/I889458895"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102996800","display_name":"Kang Min Yoo","orcid":"https://orcid.org/0000-0002-1171-6077"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kang Min Yoo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139764946","display_name":"Jeonghoon Kim","orcid":null},"institutions":[{"id":"https://openalex.org/I4405260609","display_name":"NAVER Cloud (South Korea)","ror":"https://ror.org/041051y78","country_code":"KR","type":"company","lineage":["https://openalex.org/I4405260609","https://openalex.org/I60922564"]},{"id":"https://openalex.org/I60922564","display_name":"Naver (South Korea)","ror":"https://ror.org/04nzrnx83","country_code":"KR","type":"company","lineage":["https://openalex.org/I60922564"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Jeonghoon Kim","raw_affiliation_strings":["NAVER Cloud"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NAVER Cloud","institution_ids":["https://openalex.org/I4405260609","https://openalex.org/I60922564"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5123523883","display_name":"Sungju Kim","orcid":null},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]},{"id":"https://openalex.org/I4405260609","display_name":"NAVER Cloud (South Korea)","ror":"https://ror.org/041051y78","country_code":"KR","type":"company","lineage":["https://openalex.org/I4405260609","https://openalex.org/I60922564"]},{"id":"https://openalex.org/I60922564","display_name":"Naver (South Korea)","ror":"https://ror.org/04nzrnx83","country_code":"KR","type":"company","lineage":["https://openalex.org/I60922564"]},{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK","KR"],"is_corresponding":false,"raw_author_name":"Sungju Kim","raw_affiliation_strings":["NAVER Cloud","The Hong Kong University of Science and Technology (Guangzhou )"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NAVER Cloud","institution_ids":["https://openalex.org/I4405260609","https://openalex.org/I60922564"]},{"raw_affiliation_string":"The Hong Kong University of Science and Technology (Guangzhou )","institution_ids":["https://openalex.org/I200769079","https://openalex.org/I889458895"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":4,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.76996505,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"37543","last_page":"37562"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.15320000052452087,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.15320000052452087,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12072","display_name":"Machine Learning and Algorithms","score":0.0754999965429306,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.062199998646974564,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.4950000047683716},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.41260001063346863},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.3743000030517578},{"id":"https://openalex.org/keywords/natural-language","display_name":"Natural language","score":0.28679999709129333},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.28290000557899475},{"id":"https://openalex.org/keywords/language-acquisition","display_name":"Language acquisition","score":0.26829999685287476}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7039999961853027},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.4950000047683716},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4839000105857849},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.4706000089645386},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.41260001063346863},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3743000030517578},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.37059998512268066},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.28679999709129333},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.28290000557899475},{"id":"https://openalex.org/C74672266","wikidata":"https://www.wikidata.org/wiki/Q815859","display_name":"Language acquisition","level":2,"score":0.26829999685287476},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.26260000467300415},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.26190000772476196},{"id":"https://openalex.org/C136197465","wikidata":"https://www.wikidata.org/wiki/Q1729295","display_name":"Variety (cybernetics)","level":2,"score":0.26109999418258667},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.2603999972343445},{"id":"https://openalex.org/C2777212361","wikidata":"https://www.wikidata.org/wiki/Q5127848","display_name":"Class (philosophy)","level":2,"score":0.2540000081062317},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.25110000371932983}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.findings-acl.1872","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1872","pdf_url":"https://aclanthology.org/2026.findings-acl.1872.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-acl.1872","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1872","pdf_url":"https://aclanthology.org/2026.findings-acl.1872.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166804547.pdf","grobid_xml":"https://content.openalex.org/works/W7166804547.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"While":[0],"Large":[1],"Language":[2],"Models":[3],"(LLMs)":[4],"have":[5],"revolutionized":[6],"code":[7,217],"generation,":[8,74],"standard":[9],"\"System":[10],"1\"":[11],"approaches":[12],"that":[13,66,145],"generate":[14],"solutions":[15],"in":[16,157,182],"a":[17,23,60,149,183],"single":[18],"forward":[19],"pass":[20],"often":[21],"hit":[22],"performance":[24],"ceiling":[25],"on":[26,45,131,164,169,173,176,180],"complex":[27],"algorithmic":[28],"tasks.Existing":[29],"iterative":[30],"refinement":[31,95],"strategies":[32],"attempt":[33],"to":[34,96,117,127],"bridge":[35],"this":[36,55],"gap":[37],"at":[38,105,137,222],"inference":[39,106,138],"time,":[40],"yet":[41],"they":[42],"predominantly":[43],"rely":[44],"external":[46],"oracles,":[47],"execution":[48,135],"feedback,":[49],"or":[50,134,187],"computationally":[51],"expensive":[52],"promptresponse":[53],"cycles.In":[54],"work,":[56],"we":[57],"propose":[58],"Re-flexiCoder,":[59],"novel":[61],"reinforcement":[62],"learning":[63],"(RL)":[64],"framework":[65,194],"internalizes":[67],"the":[68,84,91,119,124,158],"structured":[69],"reasoning":[70,212],"trajectory,":[71,122],"encompassing":[72],"initial":[73],"bug":[75],"and":[76,80,102,178,213,218],"optimization":[77],"aware":[78],"reflection,":[79],"self-correction,":[81],"directly":[82],"into":[83],"model's":[85],"weights.Unlike":[86],"prior":[87],"methods,":[88],"ReflexiCoder":[89],"shifts":[90],"paradigm":[92,112],"from":[93],"external-dependent":[94],"an":[97,109],"intrinsic,":[98],"fully":[99],"autonomous":[100],"self-reflection":[101],"self-correction":[103],"capabilities":[104],"time.We":[107],"utilize":[108],"RL-only":[110],"training":[111],"with":[113],"granular":[114],"reward":[115],"functions":[116],"optimize":[118],"entire":[120],"reflection-correction":[121],"teaching":[123],"model":[125],"how":[126],"debug":[128],"without":[129],"reliance":[130],"ground-truth":[132],"feedback":[133],"engines":[136],"time.Extensive":[139],"experiments":[140],"across":[141],"seven":[142],"benchmarks":[143],"demonstrate":[144],"our":[146,193],"ReflexiCoder-8B":[147],"establishes":[148],"new":[150],"state-of-the-art":[151],"(SOTA)":[152],"among":[153],"leading":[154],"open-source":[155],"models":[156,190],"1.5B-14B":[159],"range,":[160],"achieving":[161],"94.51%":[162],"(87.20%)":[163],"HumanEval":[165],"(Plus),":[166,171],"81.80%":[167],"(78.57%)":[168],"MBPP":[170],"35.00%":[172],"BigCodeBench,":[174],"52.21%":[175],"Live-CodeBench,":[177],"37.34%":[179],"CodeForces":[181],"single-attempt":[184],"setting,":[185],"rivaling":[186],"surpassing":[188],"proprietary":[189],"like":[191],"GPT-5.1.Notably,":[192],"is":[195],"significantly":[196],"more":[197],"token-efficient":[198],"than":[199],"base":[200],"models,":[201],"reducing":[202],"inference-time":[203],"compute":[204],"overhead":[205],"by":[206],"approximately":[207],"40%":[208],"through":[209],"disciplined,":[210],"efficient":[211],"reflection":[214],"patterns.The":[215],"source":[216],"data":[219],"are":[220],"available":[221],"https://github.com/juyongjiang":[223],"/ReflexiCoder.":[224]},"counts_by_year":[],"updated_date":"2026-08-21T09:56:20.448147","created_date":"2026-07-02T00:00:00"}