{"id":"https://openalex.org/W7167607092","doi":"https://doi.org/10.1145/3786181.3788714","title":"Multi-task Code LLMs: Data Mix or Model Merge?","display_name":"Multi-task Code LLMs: Data Mix or Model Merge?","publication_year":2026,"publication_date":"2026-04-12","ids":{"openalex":"https://openalex.org/W7167607092","doi":"https://doi.org/10.1145/3786181.3788714"},"language":null,"primary_location":{"id":"doi:10.1145/3786181.3788714","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3786181.3788714","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 3rd International Workshop on Large Language Models For Code","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3786181.3788714","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5136290106","display_name":"Mingzhi Zhu","orcid":null},"institutions":[{"id":"https://openalex.org/I165799507","display_name":"Rensselaer Polytechnic Institute","ror":"https://ror.org/01rtyzb94","country_code":"US","type":"education","lineage":["https://openalex.org/I165799507"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Mingzhi Zhu","raw_affiliation_strings":["Rensselaer Polytechnic Institute, Troy, NY, USA"],"raw_orcid":"https://orcid.org/0009-0000-2086-6715","affiliations":[{"raw_affiliation_string":"Rensselaer Polytechnic Institute, Troy, NY, USA","institution_ids":["https://openalex.org/I165799507"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5068061267","display_name":"Michele Merler","orcid":"https://orcid.org/0000-0002-4358-8671"},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Michele Merler","raw_affiliation_strings":["IBM Research, Yorktown Heights, USA"],"raw_orcid":"https://orcid.org/0000-0002-4358-8671","affiliations":[{"raw_affiliation_string":"IBM Research, Yorktown Heights, USA","institution_ids":["https://openalex.org/I1341412227"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5095381632","display_name":"Boris Sobolev","orcid":"https://orcid.org/0009-0009-7833-0230"},"institutions":[{"id":"https://openalex.org/I151281966","display_name":"Cisco Systems (China)","ror":"https://ror.org/02qy75381","country_code":"CN","type":"company","lineage":["https://openalex.org/I135428043","https://openalex.org/I151281966"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Boris Sobolev","raw_affiliation_strings":["Cisco, Woodbridge, NJ, USA"],"raw_orcid":"https://orcid.org/0009-0009-7833-0230","affiliations":[{"raw_affiliation_string":"Cisco, Woodbridge, NJ, USA","institution_ids":["https://openalex.org/I151281966"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5059912927","display_name":"Rahul Krishna","orcid":"https://orcid.org/0000-0002-5899-6651"},"institutions":[{"id":"https://openalex.org/I1341412227","display_name":"IBM (United States)","ror":"https://ror.org/05hh8d621","country_code":"US","type":"company","lineage":["https://openalex.org/I1341412227"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Rahul Krishna","raw_affiliation_strings":["IBM Research, Yorktown Heights, NY, USA"],"raw_orcid":"https://orcid.org/0000-0002-5899-6651","affiliations":[{"raw_affiliation_string":"IBM Research, Yorktown Heights, NY, USA","institution_ids":["https://openalex.org/I1341412227"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087072038","display_name":"Raju Pavuluri","orcid":null},"institutions":[{"id":"https://openalex.org/I4210114115","display_name":"IBM Research - Thomas J. Watson Research Center","ror":"https://ror.org/0265w5591","country_code":"US","type":"facility","lineage":["https://openalex.org/I1341412227","https://openalex.org/I4210114115"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Raju Pavuluri","raw_affiliation_strings":["IBM T.J. Watson Research Center, Yorktown Heights, NY, USA"],"raw_orcid":"https://orcid.org/0009-0008-8810-2381","affiliations":[{"raw_affiliation_string":"IBM T.J. Watson Research Center, Yorktown Heights, NY, USA","institution_ids":["https://openalex.org/I4210114115"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5081738999","display_name":"Stacy Patterson","orcid":"https://orcid.org/0000-0001-7711-6018"},"institutions":[{"id":"https://openalex.org/I165799507","display_name":"Rensselaer Polytechnic Institute","ror":"https://ror.org/01rtyzb94","country_code":"US","type":"education","lineage":["https://openalex.org/I165799507"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Stacy Patterson","raw_affiliation_strings":["Rensselaer Polytechnic Institute, Troy, NY, USA"],"raw_orcid":"https://orcid.org/0000-0001-7711-6018","affiliations":[{"raw_affiliation_string":"Rensselaer Polytechnic Institute, Troy, NY, USA","institution_ids":["https://openalex.org/I165799507"]}]}],"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":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.94677381,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"29","last_page":"36"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10260","display_name":"Software Engineering Research","score":0.4794999957084656,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10260","display_name":"Software Engineering Research","score":0.4794999957084656,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10028","display_name":"Topic Modeling","score":0.13979999721050262,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.054499998688697815,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/automatic-summarization","display_name":"Automatic summarization","score":0.8009999990463257},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.6959999799728394},{"id":"https://openalex.org/keywords/ideal","display_name":"Ideal (ethics)","score":0.44040000438690186},{"id":"https://openalex.org/keywords/scale","display_name":"Scale (ratio)","score":0.4284000098705292},{"id":"https://openalex.org/keywords/code-generation","display_name":"Code generation","score":0.40149998664855957},{"id":"https://openalex.org/keywords/data-modeling","display_name":"Data modeling","score":0.383899986743927},{"id":"https://openalex.org/keywords/software-deployment","display_name":"Software deployment","score":0.38280001282691956}],"concepts":[{"id":"https://openalex.org/C170858558","wikidata":"https://www.wikidata.org/wiki/Q1394144","display_name":"Automatic summarization","level":2,"score":0.8009999990463257},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7703999876976013},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.6959999799728394},{"id":"https://openalex.org/C2776639384","wikidata":"https://www.wikidata.org/wiki/Q840396","display_name":"Ideal (ethics)","level":2,"score":0.44040000438690186},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.4284000098705292},{"id":"https://openalex.org/C133162039","wikidata":"https://www.wikidata.org/wiki/Q1061077","display_name":"Code generation","level":3,"score":0.40149998664855957},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.383899986743927},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.38280001282691956},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3677000105381012},{"id":"https://openalex.org/C138777275","wikidata":"https://www.wikidata.org/wiki/Q6884054","display_name":"Mixing (physics)","level":2,"score":0.3321000039577484},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3264000117778778},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.31060001254081726},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3000999987125397},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.2831000089645386},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.2597000002861023},{"id":"https://openalex.org/C162319229","wikidata":"https://www.wikidata.org/wiki/Q175263","display_name":"Data structure","level":2,"score":0.25679999589920044},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.2565000057220459},{"id":"https://openalex.org/C199519371","wikidata":"https://www.wikidata.org/wiki/Q942695","display_name":"Source lines of code","level":3,"score":0.2551000118255615},{"id":"https://openalex.org/C2778598016","wikidata":"https://www.wikidata.org/wiki/Q3201279","display_name":"Code-mixing","level":3,"score":0.2533999979496002}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3786181.3788714","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3786181.3788714","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 3rd International Workshop on Large Language Models For Code","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3786181.3788714","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3786181.3788714","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 3rd International Workshop on Large Language Models For Code","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":13,"referenced_works":["https://openalex.org/W1574447377","https://openalex.org/W2758950307","https://openalex.org/W4320814215","https://openalex.org/W4401863706","https://openalex.org/W4402667093","https://openalex.org/W4403913792","https://openalex.org/W4404782653","https://openalex.org/W4404782795","https://openalex.org/W4405543707","https://openalex.org/W4406844855","https://openalex.org/W4412888517","https://openalex.org/W7133198585","https://openalex.org/W7133224126"],"related_works":[],"abstract_inverted_index":{"Recent":[0],"research":[1],"advocates":[2],"deploying":[3],"smaller,":[4],"specialized":[5,98],"code":[6,36,66,69,102,202],"LLMs":[7],"in":[8,16],"agentic":[9],"frameworks":[10],"alongside":[11],"frontier":[12],"models,":[13,117],"sparking":[14],"interest":[15],"efficient":[17],"strategies":[18,184],"for":[19,32,65,136,173,197],"multi-task":[20,35],"learning":[21],"that":[22,81,179],"balance":[23],"performance,":[24],"constraints,":[25],"and":[26,53,60,68,77,170,182],"costs.":[27],"We":[28,43,155],"compare":[29],"two":[30,48,57],"approaches":[31],"creating":[33],"small,":[34],"LLMs:":[37],"data":[38,148],"mixing":[39,149,183],"versus":[40],"model":[41,49,82,93,99,127,168],"merging.":[42],"conduct":[44],"extensive":[45],"experiments":[46],"across":[47,92],"families":[50],"(Qwen":[51],"Coder":[52,124],"DeepSeek":[54],"Coder)":[55],"at":[56,89,206],"scales":[58],"(2B":[59],"7B":[61,126],"parameters),":[62],"fine-tuning":[63],"them":[64,195],"generation":[67,103],"summarization":[70,107],"tasks.":[71],"Our":[72,201],"evaluation":[73],"on":[74,101,131],"HumanEval,":[75],"MBPP,":[76],"CodeXGlue":[78],"benchmarks":[79],"reveals":[80],"merging":[83,174,181],"achieves":[84],"the":[85],"best":[86,120],"overall":[87],"performance":[88,100,192],"larger":[90],"scale":[91,144],"families,":[94],"retaining":[95],"96%":[96],"of":[97,122],"tasks":[104,166],"while":[105],"maintaining":[106],"capabilities.":[108],"Notably,":[109],"merged":[110],"models":[111],"can":[112,185,203],"even":[113],"surpass":[114],"individually":[115],"fine-tuned":[116,139],"with":[118],"our":[119],"configuration":[121],"Qwen":[123],"2.5":[125],"achieving":[128],"92.7%":[129],"Pass@1":[130],"HumanEval":[132],"compared":[133],"to":[134,150,162],"90.9%":[135],"its":[137],"task-specific":[138,188],"equivalent.":[140],"At":[141],"a":[142,152,158],"smaller":[143],"we":[145],"find":[146],"instead":[147],"be":[151,204],"preferred":[153],"strategy.":[154],"further":[156],"introduce":[157],"weight":[159],"analysis":[160],"technique":[161],"understand":[163],"how":[164],"different":[165],"affect":[167],"parameters":[169],"their":[171],"implications":[172],"strategies.":[175],"The":[176],"results":[177],"suggest":[178],"careful":[180],"effectively":[186],"combine":[187],"capabilities":[189],"without":[190],"significant":[191],"degradation,":[193],"making":[194],"ideal":[196],"resource-constrained":[198],"deployment":[199],"scenarios.":[200],"found":[205],"https://github.com/zmzfpc/Model_Merging_Data_Mixture.":[207]},"counts_by_year":[],"updated_date":"2026-07-08T06:22:20.291053","created_date":"2026-07-08T00:00:00"}