{"id":"https://openalex.org/W7214916240","doi":"https://doi.org/10.48550/arxiv.2609.35793","title":"Learning from the Gap Between Pass@K and Pass@1","display_name":"Learning from the Gap Between Pass@K and Pass@1","publication_year":2026,"publication_date":"2026-09-17","ids":{"openalex":"https://openalex.org/W7214916240","doi":"https://doi.org/10.48550/arxiv.2609.35793"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.35793","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.35793","pdf_url":"https://arxiv.org/pdf/2609.35793","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2609.35793","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5153660592","display_name":"Xuan Liu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Xuan","raw_affiliation_strings":["University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5060291417","display_name":"Jingbin Qian","orcid":"https://orcid.org/0009-0005-4935-4568","observed_orcids":["https://orcid.org/0009-0005-4935-4568"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qian, Jingbin","raw_affiliation_strings":["University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5139916733","display_name":"Haosheng Chen","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I66867065","display_name":"East China Normal University","ror":"https://ror.org/02n96ep67","country_code":"CN","type":"education","lineage":["https://openalex.org/I66867065"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chen, Haosheng","raw_affiliation_strings":["East China Normal University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"East China Normal University","institution_ids":["https://openalex.org/I66867065"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.8622999787330627,"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":0.8622999787330627,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.04280000180006027,"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":0.029600000008940697,"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/llm-fine-tuning","display_name":"LLM fine-tuning","score":0.7590000033378601},{"id":"https://openalex.org/keywords/rejection-sampling","display_name":"rejection sampling","score":0.4790000021457672},{"id":"https://openalex.org/keywords/fault-recovery","display_name":"fault recovery","score":0.41200000047683716},{"id":"https://openalex.org/keywords/llama-3-1","display_name":"Llama 3.1","score":0.593999981880188},{"id":"https://openalex.org/keywords/mistral-7b","display_name":"Mistral-7B","score":0.7570000290870667},{"id":"https://openalex.org/keywords/post-training","display_name":"post-training","score":0.3310000002384186}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6697999835014343},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.6517000198364258},{"id":"https://openalex.org/C90329073","wikidata":"https://www.wikidata.org/wiki/Q914232","display_name":"Ask price","level":2,"score":0.6022999882698059},{"id":"https://openalex.org/C198531522","wikidata":"https://www.wikidata.org/wiki/Q485146","display_name":"Sample (material)","level":2,"score":0.5504999756813049},{"id":"https://openalex.org/C85847156","wikidata":"https://www.wikidata.org/wiki/Q59015987","display_name":"Verifiable secret sharing","level":3,"score":0.5223000049591064},{"id":"https://openalex.org/C148220186","wikidata":"https://www.wikidata.org/wiki/Q7111912","display_name":"Outcome (game theory)","level":2,"score":0.48069998621940613},{"id":"https://openalex.org/C124681953","wikidata":"https://www.wikidata.org/wiki/Q339062","display_name":"Decomposition","level":2,"score":0.4507000148296356},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.44999998807907104},{"id":"https://openalex.org/C12590798","wikidata":"https://www.wikidata.org/wiki/Q3933199","display_name":"Replication (statistics)","level":2,"score":0.4171000123023987},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.35249999165534973},{"id":"https://openalex.org/C8505890","wikidata":"https://www.wikidata.org/wiki/Q605095","display_name":"Budget constraint","level":2,"score":0.35040000081062317},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.3327000141143799},{"id":"https://openalex.org/C176248197","wikidata":"https://www.wikidata.org/wiki/Q458526","display_name":"Probably approximately correct learning","level":4,"score":0.3292999863624573},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.32749998569488525},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.2978000044822693},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.26249998807907104},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.2612999975681305},{"id":"https://openalex.org/C67203356","wikidata":"https://www.wikidata.org/wiki/Q1321905","display_name":"Reinforcement","level":2,"score":0.2590999901294708},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.2558000087738037},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.25529998540878296}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.35793","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.35793","pdf_url":"https://arxiv.org/pdf/2609.35793","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.35793","doi":"https://doi.org/10.48550/arxiv.2609.35793","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.35793","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2609.35793","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.35793","pdf_url":"https://arxiv.org/pdf/2609.35793","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7214916240.pdf","grobid_xml":"https://content.openalex.org/works/W7214916240.grobid-xml"},"referenced_works_count":17,"referenced_works":["https://openalex.org/W3210904070","https://openalex.org/W4226287673","https://openalex.org/W4383340139","https://openalex.org/W4385970164","https://openalex.org/W4386080611","https://openalex.org/W4390833131","https://openalex.org/W4391272793","https://openalex.org/W4391631327","https://openalex.org/W4401306886","https://openalex.org/W4402671620","https://openalex.org/W4403796339","https://openalex.org/W4404782827","https://openalex.org/W4405655184","https://openalex.org/W4411119300","https://openalex.org/W7131407993","https://openalex.org/W7196953391","https://openalex.org/W7196963086"],"related_works":["https://openalex.org/W4414818038","https://openalex.org/W7131407993","https://openalex.org/W7197033118","https://openalex.org/W7214807947","https://openalex.org/W7203627068","https://openalex.org/W7154239169","https://openalex.org/W7223379962","https://openalex.org/W6939131029","https://openalex.org/W6948226500","https://openalex.org/W7166822156"],"abstract_inverted_index":{"Sampling":[0],"many":[1,31,244],"responses":[2,41,117],"and":[3,88,108,112,131,143,165,168,218,236],"keeping":[4],"one":[5,97],"that":[6,91,196,202,223,237],"passes":[7],"a":[8,35,52,161],"verifier":[9],"lets":[10],"large":[11],"language":[12],"models":[13],"solve":[14],"problems":[15,62,90,134],"beyond":[16],"their":[17],"single-response":[18],"ability,":[19],"but":[20,99],"this":[21,76],"search":[22,47,206,215,231,247],"must":[23],"be":[24],"paid":[25],"again":[26],"for":[27],"every":[28,158],"query,":[29],"while":[30,209],"deployments":[32],"answer":[33],"with":[34,96,145,160,188,228],"single":[36],"response.":[37],"Post-training":[38],"on":[39,61,83,133,166,183],"verified":[40,116,186],"can":[42],"transfer":[43],"the":[44,49,63,84,92,106,124,135,179,184,197,203,225,238],"benefit":[45],"of":[46,178,181,190],"into":[48,128],"model.":[50],"With":[51],"fixed":[53,111],"budget,":[54],"selecting":[55],"by":[56,213,242],"correctness":[57],"alone":[58],"spends":[59],"slots":[60],"model":[64,94,137,147],"already":[65,138],"answers":[66],"correctly,":[67],"leaving":[68],"fewer":[69],"to":[70,175],"correct":[71],"its":[72,191],"failures.":[73],"To":[74],"address":[75],"imbalance,":[77],"we":[78],"propose":[79],"GapFT,":[80],"which":[81,115],"trains":[82],"gap":[85],"between":[86],"Pass@K":[87],"Pass@1:":[89],"source":[93,136],"fails":[95],"response":[98],"solves":[100],"within":[101],"K":[102],"samples.":[103],"GapFT":[104,149],"keeps":[105],"objective":[107],"training":[109],"budget":[110,227],"changes":[113],"only":[114,212],"enter":[118],"training;":[119],"an":[120],"exact":[121],"decomposition":[122],"splits":[123],"resulting":[125],"Pass@1":[126],"change":[127],"corrected":[129],"failures":[130,201,210,230,246],"regressions":[132],"solved.":[139],"On":[140],"LogiQA":[141],"2.0":[142],"ReClor":[144],"three":[146],"families,":[148],"is":[150,240],"above":[151],"budget-matched":[152],"uniform":[153],"rejection-sampling":[154],"fine-tuning":[155,182],"(RFT)":[156],"in":[157],"setting,":[159],"positive":[162],"pooled":[163],"effect,":[164],"Llama-3.1-8B":[167],"Mistral-7B":[169],"it":[170],"recovers":[171],"about":[172],"two":[173],"thirds":[174],"four":[176],"fifths":[177],"gain":[180,198,239],"entire":[185],"pool":[187],"11-34%":[189],"problems.":[192],"Further":[193],"analyses":[194],"reveal":[195],"comes":[199],"from":[200],"first":[204],"few":[205],"samples":[207],"recover,":[208],"found":[211],"deeper":[214],"displace":[216],"replay":[217],"add":[219],"no":[220],"net":[221],"gain,":[222],"filling":[224],"same":[226],"gold-labeled":[229],"cannot":[232],"reach":[233],"lowers":[234],"accuracy,":[235],"bounded":[241],"how":[243],"transferable":[245],"exposes.":[248]},"counts_by_year":[],"updated_date":"2026-11-08T12:12:28.789561","created_date":"2026-10-01T00:00:00"}