{"id":"https://openalex.org/W7214286160","doi":"https://doi.org/10.48550/arxiv.2609.28625","title":"RLVR landscapes for iterated multiplications can be benign: Insights from spin-glass theory","display_name":"RLVR landscapes for iterated multiplications can be benign: Insights from spin-glass theory","publication_year":2026,"publication_date":"2026-09-23","ids":{"openalex":"https://openalex.org/W7214286160","doi":"https://doi.org/10.48550/arxiv.2609.28625"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.28625","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.28625","pdf_url":"https://arxiv.org/pdf/2609.28625","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/2609.28625","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5108309346","display_name":"Noa Rubin","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I197251160","display_name":"Hebrew University of Jerusalem","ror":"https://ror.org/03qxff017","country_code":"IL","type":"education","lineage":["https://openalex.org/I197251160"]}],"countries":["IL"],"is_corresponding":false,"raw_author_name":"Rubin, Noa","raw_affiliation_strings":["Racah Institute of Physics The Hebrew University of Jerusalem"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Racah Institute of Physics The Hebrew University of Jerusalem","institution_ids":["https://openalex.org/I197251160"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5053390682","display_name":"Zohar Ringel","orcid":"https://orcid.org/0000-0002-2555-7695","observed_orcids":["https://orcid.org/0000-0002-2555-7695"]},"institutions":[{"id":"https://openalex.org/I197251160","display_name":"Hebrew University of Jerusalem","ror":"https://ror.org/03qxff017","country_code":"IL","type":"education","lineage":["https://openalex.org/I197251160"]}],"countries":["IL"],"is_corresponding":false,"raw_author_name":"Ringel, Zohar","raw_affiliation_strings":["Racah Institute of Physics The Hebrew University of Jerusalem"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Racah Institute of Physics The Hebrew University of Jerusalem","institution_ids":["https://openalex.org/I197251160"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I197251160"],"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.9789999723434448,"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":0.9789999723434448,"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":0.006399999838322401,"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/T10028","display_name":"Topic Modeling","score":0.0010999999940395355,"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/reinforcement-learning-with-verifiable-rewards","display_name":"reinforcement learning with verifiable rewards","score":0.9959999918937683},{"id":"https://openalex.org/keywords/optimization-landscape","display_name":"optimization landscape","score":0.8090000152587891},{"id":"https://openalex.org/keywords/spin-glass-theory","display_name":"spin glass theory","score":0.8489999771118164},{"id":"https://openalex.org/keywords/energy-based-models","display_name":"energy-based models","score":0.7789999842643738},{"id":"https://openalex.org/keywords/entropy-regularization","display_name":"entropy regularization","score":0.718999981880188},{"id":"https://openalex.org/keywords/local-minima","display_name":"local minima","score":0.7239999771118164},{"id":"https://openalex.org/keywords/terrain-ruggedness","display_name":"terrain ruggedness","score":0.5389999747276306},{"id":"https://openalex.org/keywords/group-multiplication","display_name":"group multiplication","score":0.6579999923706055},{"id":"https://openalex.org/keywords/non-abelian-groups","display_name":"non-abelian groups","score":0.9010000228881836},{"id":"https://openalex.org/keywords/algorithmic-reasoning","display_name":"algorithmic reasoning","score":0.7390000224113464}],"concepts":[{"id":"https://openalex.org/C140479938","wikidata":"https://www.wikidata.org/wiki/Q5254619","display_name":"Iterated function","level":2,"score":0.6510999798774719},{"id":"https://openalex.org/C186633575","wikidata":"https://www.wikidata.org/wiki/Q845060","display_name":"Maxima and minima","level":2,"score":0.6011999845504761},{"id":"https://openalex.org/C176809094","wikidata":"https://www.wikidata.org/wiki/Q15401496","display_name":"Traverse","level":2,"score":0.6008999943733215},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5623999834060669},{"id":"https://openalex.org/C85847156","wikidata":"https://www.wikidata.org/wiki/Q59015987","display_name":"Verifiable secret sharing","level":3,"score":0.5550000071525574},{"id":"https://openalex.org/C106301342","wikidata":"https://www.wikidata.org/wiki/Q4117933","display_name":"Entropy (arrow of time)","level":2,"score":0.5317000150680542},{"id":"https://openalex.org/C169345407","wikidata":"https://www.wikidata.org/wiki/Q8216221","display_name":"Uncorrelated","level":2,"score":0.5271000266075134},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.47380000352859497},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.4546000063419342},{"id":"https://openalex.org/C2777212361","wikidata":"https://www.wikidata.org/wiki/Q5127848","display_name":"Class (philosophy)","level":2,"score":0.4368000030517578},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.37959998846054077},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.36820000410079956},{"id":"https://openalex.org/C2781311116","wikidata":"https://www.wikidata.org/wiki/Q83306","display_name":"Group (periodic table)","level":2,"score":0.34549999237060547},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3312999904155731},{"id":"https://openalex.org/C119621388","wikidata":"https://www.wikidata.org/wiki/Q5377166","display_name":"Energy landscape","level":2,"score":0.3018999993801117},{"id":"https://openalex.org/C9679016","wikidata":"https://www.wikidata.org/wiki/Q1417473","display_name":"Principle of maximum entropy","level":2,"score":0.29490000009536743},{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.257999986410141},{"id":"https://openalex.org/C22367795","wikidata":"https://www.wikidata.org/wiki/Q7625208","display_name":"Structured prediction","level":2,"score":0.25459998846054077}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.28625","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.28625","pdf_url":"https://arxiv.org/pdf/2609.28625","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.2609.28625","doi":"https://doi.org/10.48550/arxiv.2609.28625","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.28625","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:2609.28625","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.28625","pdf_url":"https://arxiv.org/pdf/2609.28625","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/W7214286160.pdf","grobid_xml":"https://content.openalex.org/works/W7214286160.grobid-xml"},"referenced_works_count":20,"referenced_works":["https://openalex.org/W2111471293","https://openalex.org/W2225156818","https://openalex.org/W2799151646","https://openalex.org/W2948432982","https://openalex.org/W3039845099","https://openalex.org/W4280502627","https://openalex.org/W4360892090","https://openalex.org/W4414281281","https://openalex.org/W4414898977","https://openalex.org/W4415097709","https://openalex.org/W4416048648","https://openalex.org/W4416157008","https://openalex.org/W4417167303","https://openalex.org/W7083334705","https://openalex.org/W7127319615","https://openalex.org/W7127658205","https://openalex.org/W7160899677","https://openalex.org/W7164047390","https://openalex.org/W7196953391","https://openalex.org/W7196963086"],"related_works":["https://openalex.org/W7151366016","https://openalex.org/W7171321448","https://openalex.org/W7129660610","https://openalex.org/W7119523285","https://openalex.org/W7214035531","https://openalex.org/W4417531499","https://openalex.org/W4417113061","https://openalex.org/W7204921312","https://openalex.org/W7128466166","https://openalex.org/W4416241498"],"abstract_inverted_index":{"Despite":[0],"the":[1,10,25,71,109,133,152,167],"importance":[2],"of":[3,28,88,112,169,193],"reinforcement":[4],"learning":[5],"with":[6,92,173],"verifiable":[7],"rewards":[8],"(RLVR),":[9],"extent":[11],"to":[12,116],"which":[13],"it":[14],"can":[15,64,140,162],"learn":[16],"new":[17],"reasoning":[18],"capabilities":[19],"remains":[20],"debated.":[21],"Here":[22],"we":[23,43,176],"study":[24],"optimization":[26],"landscape":[27,72,96,153],"RLVR":[29,46,63,106],"on":[30],"algorithmic":[31,191],"tasks,":[32],"such":[33,124],"as":[34,125],"iterated":[35,196],"group":[36,198],"and":[37,66,81,90,128],"quasigroup":[38],"multiplication.":[39],"To":[40],"this":[41,74,95,174],"end,":[42],"map":[44],"entropy-regularized":[45],"over":[47,56],"myopic":[48],"tabular":[49,75],"policies":[50],"onto":[51],"an":[52,190],"energy-based":[53],"(spin-glass)":[54],"model":[55],"deterministic":[57],"policies.":[58],"This":[59],"mapping":[60],"upper-bounds":[61],"what":[62],"achieve,":[65],"lets":[67],"us":[68],"rigorously":[69],"characterize":[70],"in":[73,120,131],"setting.":[76],"We":[77,157],"show,":[78],"both":[79],"theoretically":[80],"experimentally,":[82],"that":[83,103,139,159,178],"for":[84,195],"a":[85,142,179],"wide":[86],"class":[87],"models":[89],"tasks":[91,114],"uncorrelated":[93],"inputs,":[94],"is":[97],"benign,":[98],"containing":[99],"no":[100],"local":[101],"minima":[102],"could":[104],"trap":[105],"training.":[107],"Rather,":[108],"practical":[110],"difficulty":[111],"these":[113,160],"appears":[115],"stem,":[117],"at":[118],"least":[119],"part,":[121],"from":[122,144,151,182],"issues":[123],"diffusive":[126],"barriers":[127],"gradient-estimation":[129],"error":[130],"traversing":[132],"landscape.":[134],"These":[135],"are":[136,149],"genuine":[137],"obstacles":[138,161],"prevent":[141],"solution":[143],"being":[145,155],"found,":[146],"but":[147],"they":[148],"distinct":[150],"itself":[154],"rugged.":[156],"show":[158],"often":[163],"be":[164],"mitigated":[165],"through":[166],"choice":[168],"entropy":[170],"regulator.":[171],"Consistent":[172],"theory,":[175],"find":[177],"transformer":[180],"trained":[181],"scratch,":[183],"using":[184],"only":[185],"last-token":[186],"rewards,":[187],"successfully":[188],"learns":[189],"chain":[192],"thought":[194],"non-Abelian":[197],"multiplications.":[199]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-09-26T00:00:00"}