{"id":"https://openalex.org/W7154765200","doi":"https://doi.org/10.48550/arxiv.2604.14877","title":"Does RL Expand the Capability Boundary of LLM Agents? A PASS@(k,T) Analysis","display_name":"Does RL Expand the Capability Boundary of LLM Agents? A PASS@(k,T) Analysis","publication_year":2026,"publication_date":"2026-04-16","ids":{"openalex":"https://openalex.org/W7154765200","doi":"https://doi.org/10.48550/arxiv.2604.14877"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.14877","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.14877","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2604.14877","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5068904766","display_name":"Zhiyuan Zhai","orcid":"https://orcid.org/0000-0003-1930-9633"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhai, Zhiyuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133902180","display_name":"Wenjing Yan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yan, Wenjing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133837621","display_name":"Xiaodan Shao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shao, Xiaodan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5133905305","display_name":"Xin Eric Wang","orcid":"https://orcid.org/0009-0005-0216-8661"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Xin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.7487999796867371,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.7487999796867371,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.11540000140666962,"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/T10653","display_name":"Robot Manipulation and Learning","score":0.015300000086426735,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.6656000018119812},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.6656000018119812},{"id":"https://openalex.org/keywords/boundary","display_name":"Boundary (topology)","score":0.6444000005722046},{"id":"https://openalex.org/keywords/metric","display_name":"Metric (unit)","score":0.6420000195503235},{"id":"https://openalex.org/keywords/base","display_name":"Base (topology)","score":0.6291000247001648},{"id":"https://openalex.org/keywords/sampling","display_name":"Sampling (signal processing)","score":0.40119999647140503},{"id":"https://openalex.org/keywords/work","display_name":"Work (physics)","score":0.3894999921321869},{"id":"https://openalex.org/keywords/performance-metric","display_name":"Performance metric","score":0.3822999894618988},{"id":"https://openalex.org/keywords/knowledge-base","display_name":"Knowledge base","score":0.37700000405311584}],"concepts":[{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.6656000018119812},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.6656000018119812},{"id":"https://openalex.org/C62354387","wikidata":"https://www.wikidata.org/wiki/Q875399","display_name":"Boundary (topology)","level":2,"score":0.6444000005722046},{"id":"https://openalex.org/C176217482","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Metric (unit)","level":2,"score":0.6420000195503235},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6341000199317932},{"id":"https://openalex.org/C42058472","wikidata":"https://www.wikidata.org/wiki/Q810214","display_name":"Base (topology)","level":2,"score":0.6291000247001648},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4196000099182129},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.40119999647140503},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.3894999921321869},{"id":"https://openalex.org/C2780898871","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Performance metric","level":2,"score":0.3822999894618988},{"id":"https://openalex.org/C4554734","wikidata":"https://www.wikidata.org/wiki/Q593744","display_name":"Knowledge base","level":2,"score":0.37700000405311584},{"id":"https://openalex.org/C89611455","wikidata":"https://www.wikidata.org/wiki/Q6804646","display_name":"Mechanism (biology)","level":2,"score":0.3450999855995178},{"id":"https://openalex.org/C167723999","wikidata":"https://www.wikidata.org/wiki/Q3773214","display_name":"Sampling distribution","level":2,"score":0.32580000162124634},{"id":"https://openalex.org/C110121322","wikidata":"https://www.wikidata.org/wiki/Q865811","display_name":"Distribution (mathematics)","level":2,"score":0.29840001463890076},{"id":"https://openalex.org/C176809094","wikidata":"https://www.wikidata.org/wiki/Q15401496","display_name":"Traverse","level":2,"score":0.2849000096321106},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.28450000286102295},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.2741999924182892},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.2721000015735626},{"id":"https://openalex.org/C48372109","wikidata":"https://www.wikidata.org/wiki/Q3913","display_name":"Binary number","level":2,"score":0.26969999074935913},{"id":"https://openalex.org/C67203356","wikidata":"https://www.wikidata.org/wiki/Q1321905","display_name":"Reinforcement","level":2,"score":0.26829999685287476},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.2680000066757202},{"id":"https://openalex.org/C61797465","wikidata":"https://www.wikidata.org/wiki/Q1188986","display_name":"Term (time)","level":2,"score":0.26649999618530273},{"id":"https://openalex.org/C2781018962","wikidata":"https://www.wikidata.org/wiki/Q5164884","display_name":"Container (type theory)","level":2,"score":0.26339998841285706},{"id":"https://openalex.org/C2778775528","wikidata":"https://www.wikidata.org/wiki/Q5135432","display_name":"Closing (real estate)","level":2,"score":0.25600001215934753},{"id":"https://openalex.org/C113336015","wikidata":"https://www.wikidata.org/wiki/Q574010","display_name":"Complete information","level":2,"score":0.25189998745918274}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.14877","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.14877","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2604.14877","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.14877","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Does":[0],"reinforcement":[1],"learning":[2],"genuinely":[3,88],"expand":[4],"what":[5],"LLM":[6],"agents":[7],"can":[8],"do,":[9],"or":[10],"merely":[11],"make":[12],"them":[13],"more":[14,166],"reliable?":[15],"For":[16],"static":[17],"reasoning,":[18],"recent":[19],"work":[20,128],"answers":[21],"the":[22,83,90,93,99,103,137,140,148,156,161,173,178],"second:":[23],"base":[24,100,157],"and":[25,66,102,187],"RL":[26,87,94,124,154,191],"pass@k":[27],"curves":[28],"converge":[29],"at":[30,106],"large":[31,107],"k.":[32],"We":[33,54],"ask":[34],"whether":[35],"this":[36],"holds":[37],"for":[38,192],"agentic":[39],"tool":[40],"use,":[41],"where":[42],"T":[43],"rounds":[44],"of":[45,190],"interaction":[46,67],"enable":[47],"compositional":[48,142],"strategies":[49],"that":[50,60],"re-sampling":[51],"cannot":[52],"recover.":[53],"introduce":[55],"PASS@(k,T),":[56],"a":[57,169],"two-dimensional":[58],"metric":[59],"jointly":[61],"varies":[62],"sampling":[63],"budget":[64],"k":[65,108],"depth":[68],"T,":[69],"separating":[70],"capability":[71,91],"expansion":[72,113],"from":[73],"efficiency":[74],"improvement.":[75],"Our":[76],"main":[77],"finding":[78],"is":[79,114],"that,":[80],"contrary":[81],"to":[82,116],"static-reasoning":[84],"result,":[85],"tool-use":[86],"enlarges":[89],"boundary:":[92],"agent's":[95],"pass-curve":[96],"pulls":[97],"above":[98],"model's":[101],"gap":[104],"widens":[105],"rather":[109],"than":[110],"converging.":[111],"The":[112],"specific":[115],"compositional,":[117],"sequential":[118],"information":[119],"gathering;":[120],"on":[121,139,176,197],"simpler":[122],"tasks":[123],"behaves":[125],"as":[126,147],"prior":[127],"predicts.":[129],"Under":[130],"matched":[131],"training":[132],"data,":[133],"supervised":[134],"fine-tuning":[135],"regresses":[136],"boundary":[138],"same":[141],"tasks,":[143],"isolating":[144],"self-directed":[145],"exploration":[146],"causal":[149],"factor.":[150],"Mechanism":[151],"analysis":[152],"shows":[153],"reweights":[155],"strategy":[158],"distribution":[159],"toward":[160],"subset":[162],"whose":[163],"downstream":[164],"reasoning":[165],"often":[167],"yields":[168],"correct":[170],"answer,":[171],"with":[172],"improvement":[174],"concentrated":[175],"how":[177],"agent":[179],"integrates":[180],"retrieved":[181],"information.":[182],"These":[183],"results":[184],"reconcile":[185],"optimistic":[186],"pessimistic":[188],"readings":[189],"LLMs:":[193],"both":[194],"are":[195],"correct,":[196],"different":[198],"task":[199],"types.":[200]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-04-18T00:00:00"}