{"id":"https://openalex.org/W4415318358","doi":"https://doi.org/10.48550/arxiv.2510.07312","title":"h1: Bootstrapping LLMs to Reason over Longer Horizons via Reinforcement Learning","display_name":"h1: Bootstrapping LLMs to Reason over Longer Horizons via Reinforcement Learning","publication_year":2025,"publication_date":"2025-10-08","ids":{"openalex":"https://openalex.org/W4415318358","doi":"https://doi.org/10.48550/arxiv.2510.07312"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2510.07312","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2510.07312","pdf_url":"https://arxiv.org/pdf/2510.07312","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":"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/2510.07312","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5092964228","display_name":"Sumeet Motwani","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Motwani, Sumeet Ramesh","raw_affiliation_strings":["University of Oxford"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Oxford","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5022459043","display_name":"A. V. Ivanova","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Ivanova, Alesia","raw_affiliation_strings":["University of Oxford"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Oxford","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Cai, Ziyang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I20089843","display_name":"Princeton University","ror":"https://ror.org/00hx57361","country_code":"US","type":"education","lineage":["https://openalex.org/I20089843"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Cai, Ziyang","raw_affiliation_strings":["Princeton University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Princeton University","institution_ids":["https://openalex.org/I20089843"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5001534492","display_name":"Philip H. S. Torr","orcid":"https://orcid.org/0009-0006-0259-5732","observed_orcids":["https://orcid.org/0009-0006-0259-5732"]},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Torr, Philip","raw_affiliation_strings":["University of Oxford"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Oxford","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5079926596","display_name":"Riashat Islam","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Islam, Riashat","raw_affiliation_strings":["Microsoft AI Frontiers"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft AI Frontiers","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101671828","display_name":"Shital Shah","orcid":"https://orcid.org/0000-0002-8995-1704","observed_orcids":["https://orcid.org/0000-0002-8995-1704"]},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Shah, Shital","raw_affiliation_strings":["Microsoft AI Frontiers"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft AI Frontiers","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":null,"display_name":"de Witt, Christian Schroeder","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"de Witt, Christian Schroeder","raw_affiliation_strings":["University of Oxford"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Oxford","institution_ids":["https://openalex.org/I40120149"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5129682583","display_name":"Charles London","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I40120149","display_name":"University of Oxford","ror":"https://ror.org/052gg0110","country_code":"GB","type":"education","lineage":["https://openalex.org/I40120149"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"London, Charles","raw_affiliation_strings":["University of Oxford"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Oxford","institution_ids":["https://openalex.org/I40120149"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":10,"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":0.42579999566078186,"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":0.42579999566078186,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.13289999961853027,"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":0.07039999961853027,"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/long-horizon-reasoning","display_name":"long-horizon reasoning","score":0.9919999837875366},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.8259999752044678},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"reinforcement learning","score":0.597000002861023},{"id":"https://openalex.org/keywords/curriculum-learning","display_name":"curriculum learning","score":0.8659999966621399},{"id":"https://openalex.org/keywords/sample-complexity","display_name":"sample complexity","score":0.6060000061988831}],"concepts":[{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.6601999998092651},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6126000285148621},{"id":"https://openalex.org/C207609745","wikidata":"https://www.wikidata.org/wiki/Q4944086","display_name":"Bootstrapping (finance)","level":2,"score":0.6060000061988831},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5856999754905701},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.5199000239372253},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5156000256538391},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.48089998960494995},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.43549999594688416},{"id":"https://openalex.org/C198531522","wikidata":"https://www.wikidata.org/wiki/Q485146","display_name":"Sample (material)","level":2,"score":0.4163999855518341},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.4074000120162964},{"id":"https://openalex.org/C47177190","wikidata":"https://www.wikidata.org/wiki/Q207137","display_name":"Curriculum","level":2,"score":0.4020000100135803},{"id":"https://openalex.org/C159176650","wikidata":"https://www.wikidata.org/wiki/Q43261","display_name":"Horizon","level":2,"score":0.36000001430511475},{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.3441999852657318},{"id":"https://openalex.org/C2778445095","wikidata":"https://www.wikidata.org/wiki/Q18354077","display_name":"Sample complexity","level":2,"score":0.3181000053882599},{"id":"https://openalex.org/C19768560","wikidata":"https://www.wikidata.org/wiki/Q320727","display_name":"Dependency (UML)","level":2,"score":0.30809998512268066},{"id":"https://openalex.org/C148220186","wikidata":"https://www.wikidata.org/wiki/Q7111912","display_name":"Outcome (game theory)","level":2,"score":0.2962999939918518},{"id":"https://openalex.org/C151376022","wikidata":"https://www.wikidata.org/wiki/Q168698","display_name":"Exponential function","level":2,"score":0.29350000619888306},{"id":"https://openalex.org/C107673813","wikidata":"https://www.wikidata.org/wiki/Q812534","display_name":"Bayesian probability","level":2,"score":0.29010000824928284},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.2816999852657318},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.26669999957084656},{"id":"https://openalex.org/C28761237","wikidata":"https://www.wikidata.org/wiki/Q7805321","display_name":"Time horizon","level":2,"score":0.2646999955177307},{"id":"https://openalex.org/C2777508537","wikidata":"https://www.wikidata.org/wiki/Q7936620","display_name":"Visual reasoning","level":2,"score":0.260699987411499},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.25940001010894775},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.2524000108242035}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2510.07312","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2510.07312","pdf_url":"https://arxiv.org/pdf/2510.07312","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2510.07312","doi":"https://doi.org/10.48550/arxiv.2510.07312","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2510.07312","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:2510.07312","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2510.07312","pdf_url":"https://arxiv.org/pdf/2510.07312","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4415318358.pdf","grobid_xml":"https://content.openalex.org/works/W4415318358.grobid-xml"},"referenced_works_count":31,"referenced_works":["https://openalex.org/W2119717200","https://openalex.org/W2192240806","https://openalex.org/W2984272647","https://openalex.org/W3093356033","https://openalex.org/W3134642945","https://openalex.org/W4391871437","https://openalex.org/W4393942918","https://openalex.org/W4399408369","https://openalex.org/W4402733601","https://openalex.org/W4403324134","https://openalex.org/W4406231668","https://openalex.org/W4407184544","https://openalex.org/W4407370164","https://openalex.org/W4409049124","https://openalex.org/W4414281281","https://openalex.org/W4414859475","https://openalex.org/W4415057368","https://openalex.org/W4415796274","https://openalex.org/W4415798713","https://openalex.org/W4416036482","https://openalex.org/W4416048648","https://openalex.org/W4416613006","https://openalex.org/W4417095241","https://openalex.org/W4417170767","https://openalex.org/W4417258158","https://openalex.org/W4417357116","https://openalex.org/W6903651657","https://openalex.org/W7083683602","https://openalex.org/W7083820855","https://openalex.org/W7133192525","https://openalex.org/W7196963086"],"related_works":["https://openalex.org/W4416383416","https://openalex.org/W7127567581","https://openalex.org/W4415309319","https://openalex.org/W4417119453","https://openalex.org/W7222994624","https://openalex.org/W7103167281","https://openalex.org/W7160633966","https://openalex.org/W4414859475","https://openalex.org/W4407231747","https://openalex.org/W6891927978"],"abstract_inverted_index":{"Large":[0],"language":[1],"models":[2,69,152],"excel":[3],"at":[4,147],"short-horizon":[5,51],"reasoning":[6,12,45,156],"tasks,":[7],"but":[8],"performance":[9],"drops":[10],"as":[11],"horizon":[13],"lengths":[14],"increase.":[15],"Existing":[16],"approaches":[17],"to":[18,42,87,120,126,183],"combat":[19],"this":[20,35,71],"rely":[21],"on":[22,70,102,111],"inference-time":[23],"scaffolding":[24],"or":[25],"costly":[26],"step-level":[27],"supervision,":[28],"neither":[29],"of":[30,64],"which":[31],"scales":[32],"easily.":[33],"In":[34],"work,":[36],"we":[37,161],"introduce":[38],"a":[39,77],"scalable":[40],"method":[41,96],"bootstrap":[43],"long-horizon":[44,139,196],"capabilities":[46],"using":[47,73,198],"only":[48,199],"existing,":[49],"abundant":[50],"data.":[52,201],"Our":[53],"approach":[54],"synthetically":[55],"composes":[56],"simple":[57],"problems":[58,107,197],"into":[59],"complex,":[60],"multi-step":[61],"dependency":[62],"chains":[63],"arbitrary":[65],"length.":[66],"We":[67],"train":[68],"data":[72],"outcome-only":[74],"rewards":[75,168],"under":[76,158],"curriculum":[78,100,164],"that":[79,151,163],"automatically":[80],"increases":[81],"in":[82,173],"complexity,":[83],"allowing":[84],"RL":[85,165,194],"training":[86,101,180],"be":[88],"scaled":[89],"much":[90],"further":[91],"without":[92],"saturating.":[93],"Empirically,":[94],"our":[95,138],"generalizes":[97],"remarkably":[98],"well:":[99],"composed":[103],"6th-grade":[104],"level":[105],"math":[106],"(GSM8K)":[108],"boosts":[109],"accuracy":[110],"longer,":[112],"competition-level":[113],"benchmarks":[114],"(GSM-Symbolic,":[115],"MATH-500,":[116],"AIME)":[117],"by":[118],"up":[119],"2.06x.":[121],"It":[122],"also":[123],"transfers":[124],"significantly":[125,142],"diverse":[127],"out-of-distribution":[128],"ReasoningGym":[129],"domains":[130],"and":[131],"long-context":[132],"benchmarks,":[133],"indicating":[134],"broader":[135],"generalization.":[136],"Importantly,":[137],"improvements":[140],"are":[141],"higher":[143],"than":[144],"baselines":[145],"even":[146],"high":[148],"pass@k,":[149],"showing":[150],"can":[153],"learn":[154],"new":[155],"paths":[157],"RL.":[159],"Theoretically,":[160],"show":[162],"with":[166],"outcome":[167],"achieves":[169],"an":[170,189],"exponential":[171],"improvement":[172],"sample":[174],"complexity":[175],"over":[176],"full-horizon":[177],"training,":[178],"providing":[179],"signal":[181],"comparable":[182],"dense":[184],"supervision.":[185],"h1":[186],"therefore":[187],"introduces":[188],"efficient":[190],"path":[191],"towards":[192],"scaling":[193],"for":[195],"existing":[200]},"counts_by_year":[{"year":2026,"cited_by_count":10}],"updated_date":"2026-11-09T13:22:03.289127","created_date":"2025-10-18T00:00:00"}