{"id":"https://openalex.org/W7204251437","doi":"https://doi.org/10.48550/arxiv.2608.24136","title":"Steering Recurrent Reasoners at Inference Time with Readout Feedback","display_name":"Steering Recurrent Reasoners at Inference Time with Readout Feedback","publication_year":2026,"publication_date":"2026-08-25","ids":{"openalex":"https://openalex.org/W7204251437","doi":"https://doi.org/10.48550/arxiv.2608.24136"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.24136","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.24136","pdf_url":"https://arxiv.org/pdf/2608.24136","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","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2608.24136","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5133744305","display_name":"Shunsuke Kamiya","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I74801974","display_name":"The University of Tokyo","ror":"https://ror.org/057zh3y96","country_code":"JP","type":"education","lineage":["https://openalex.org/I74801974"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Kamiya, Shunsuke","raw_affiliation_strings":["Graduate School of Engineering , The University of Tokyo"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Graduate School of Engineering , The University of Tokyo","institution_ids":["https://openalex.org/I74801974"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103751566","display_name":"Masanori Koyama","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I74801974","display_name":"The University of Tokyo","ror":"https://ror.org/057zh3y96","country_code":"JP","type":"education","lineage":["https://openalex.org/I74801974"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Koyama, Masanori","raw_affiliation_strings":["Graduate School of Engineering , The University of Tokyo"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Graduate School of Engineering , The University of Tokyo","institution_ids":["https://openalex.org/I74801974"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148177562","display_name":"Seongcheol Jeong","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I74801974","display_name":"The University of Tokyo","ror":"https://ror.org/057zh3y96","country_code":"JP","type":"education","lineage":["https://openalex.org/I74801974"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Jeong, Seongcheol","raw_affiliation_strings":["Graduate School of Engineering , The University of Tokyo"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Graduate School of Engineering , The University of Tokyo","institution_ids":["https://openalex.org/I74801974"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5079445397","display_name":"Fumiya Uchiyama","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I74801974","display_name":"The University of Tokyo","ror":"https://ror.org/057zh3y96","country_code":"JP","type":"education","lineage":["https://openalex.org/I74801974"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Uchiyama, Fumiya","raw_affiliation_strings":["Graduate School of Engineering , The University of Tokyo"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Graduate School of Engineering , The University of Tokyo","institution_ids":["https://openalex.org/I74801974"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133776228","display_name":"Kenji Kubo","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I74801974","display_name":"The University of Tokyo","ror":"https://ror.org/057zh3y96","country_code":"JP","type":"education","lineage":["https://openalex.org/I74801974"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Kubo, Kenji","raw_affiliation_strings":["Graduate School of Engineering , The University of Tokyo"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Graduate School of Engineering , The University of Tokyo","institution_ids":["https://openalex.org/I74801974"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5132867771","display_name":"Kohei Hayashi","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I74801974","display_name":"The University of Tokyo","ror":"https://ror.org/057zh3y96","country_code":"JP","type":"education","lineage":["https://openalex.org/I74801974"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Hayashi, Kohei","raw_affiliation_strings":["Graduate School of Engineering , The University of Tokyo"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Graduate School of Engineering , The University of Tokyo","institution_ids":["https://openalex.org/I74801974"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139985446","display_name":"Masahiro Suzuki","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I74801974","display_name":"The University of Tokyo","ror":"https://ror.org/057zh3y96","country_code":"JP","type":"education","lineage":["https://openalex.org/I74801974"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Suzuki, Masahiro","raw_affiliation_strings":["Graduate School of Engineering , The University of Tokyo"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Graduate School of Engineering , The University of Tokyo","institution_ids":["https://openalex.org/I74801974"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5129787165","display_name":"Yutaka Matsuo","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I74801974","display_name":"The University of Tokyo","ror":"https://ror.org/057zh3y96","country_code":"JP","type":"education","lineage":["https://openalex.org/I74801974"]}],"countries":["JP"],"is_corresponding":false,"raw_author_name":"Matsuo, Yutaka","raw_affiliation_strings":["Graduate School of Engineering , The University of Tokyo"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Graduate School of Engineering , The University of Tokyo","institution_ids":["https://openalex.org/I74801974"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I74801974"],"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/T10906","display_name":"AI-based Problem Solving and Planning","score":0.3125,"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/T10906","display_name":"AI-based Problem Solving and Planning","score":0.3125,"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":0.2313999980688095,"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/T11010","display_name":"Logic, Reasoning, and Knowledge","score":0.125900000333786,"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/inference-time-control","display_name":"inference-time control","score":0.796999990940094},{"id":"https://openalex.org/keywords/feedback-control","display_name":"feedback control","score":0.6349999904632568},{"id":"https://openalex.org/keywords/trm","display_name":"TRM","score":0.9610000252723694},{"id":"https://openalex.org/keywords/sudoku-solving","display_name":"Sudoku solving","score":0.703000009059906},{"id":"https://openalex.org/keywords/maze-solving","display_name":"maze solving","score":0.9200000166893005},{"id":"https://openalex.org/keywords/computational-efficiency","display_name":"computational efficiency","score":0.503000020980835}],"concepts":[{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.7551000118255615},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.6915000081062317},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6779999732971191},{"id":"https://openalex.org/C184898388","wikidata":"https://www.wikidata.org/wiki/Q1435712","display_name":"Pairwise comparison","level":2,"score":0.6269000172615051},{"id":"https://openalex.org/C131584629","wikidata":"https://www.wikidata.org/wiki/Q4308705","display_name":"Coupling (piping)","level":2,"score":0.44119998812675476},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4408999979496002},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.4260999858379364},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.37400001287460327},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.36309999227523804},{"id":"https://openalex.org/C89611455","wikidata":"https://www.wikidata.org/wiki/Q6804646","display_name":"Mechanism (biology)","level":2,"score":0.34869998693466187},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3425999879837036},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.334199994802475},{"id":"https://openalex.org/C179799912","wikidata":"https://www.wikidata.org/wiki/Q205084","display_name":"Computational complexity theory","level":2,"score":0.3206000030040741},{"id":"https://openalex.org/C145912823","wikidata":"https://www.wikidata.org/wiki/Q113558","display_name":"Dynamics (music)","level":2,"score":0.30550000071525574},{"id":"https://openalex.org/C2778445095","wikidata":"https://www.wikidata.org/wiki/Q18354077","display_name":"Sample complexity","level":2,"score":0.3034000098705292},{"id":"https://openalex.org/C51167844","wikidata":"https://www.wikidata.org/wiki/Q4422623","display_name":"Latent variable","level":2,"score":0.28949999809265137},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.28540000319480896},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.25839999318122864}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.24136","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.24136","pdf_url":"https://arxiv.org/pdf/2608.24136","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.2608.24136","doi":"https://doi.org/10.48550/arxiv.2608.24136","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.24136","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:2608.24136","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.24136","pdf_url":"https://arxiv.org/pdf/2608.24136","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/W7204251437.pdf","grobid_xml":"https://content.openalex.org/works/W7204251437.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7214986085","https://openalex.org/W7221186368","https://openalex.org/W7162411103","https://openalex.org/W7214904958","https://openalex.org/W7208734415","https://openalex.org/W7128371330","https://openalex.org/W4417100330","https://openalex.org/W7204222467","https://openalex.org/W7166676514","https://openalex.org/W7168030527"],"abstract_inverted_index":{"Recurrent":[0],"models,":[1],"which":[2],"repeatedly":[3],"update":[4],"latent":[5,61,85,144],"states":[6],"with":[7],"shared":[8],"computation":[9,25],"blocks,":[10],"have":[11],"emerged":[12],"as":[13,148],"powerful":[14],"architectures":[15],"for":[16,154],"solving":[17],"complex":[18],"reasoning":[19,156],"tasks.":[20],"Existing":[21],"inference-time":[22,151],"methods":[23],"scale":[24],"by":[26,53,121],"running":[27,123],"more":[28,32,124],"steps":[29,125],"or":[30,126,133],"sampling":[31],"trajectories,":[33,130],"but":[34],"ignore":[35],"information":[36],"revealed":[37],"within":[38],"each":[39],"trajectory.":[40],"Here":[41],"we":[42],"show":[43],"that":[44,73,140],"recurrent":[45,89,155],"models":[46,90],"can":[47,146],"be":[48],"improved":[49],"at":[50,131],"inference":[51],"time":[52],"using":[54],"their":[55],"own":[56],"readout":[57],"probabilities":[58],"to":[59],"steer":[60],"dynamics":[62,145],"without":[63],"retraining.":[64],"We":[65],"introduce":[66],"Readout":[67],"Feedback":[68],"(RoFB),":[69],"a":[70,109,149],"test-time":[71],"intervention":[72],"converts":[74],"intermediate":[75],"predictions":[76],"into":[77,83],"token-wise":[78],"pairwise":[79],"coupling":[80],"forces":[81],"injected":[82],"the":[84],"dynamics.":[86],"Across":[87],"three":[88],"(AKOrN,":[91],"ItrSA++,":[92],"TRM)":[93],"on":[94],"Sudoku":[95],"and":[96,108],"Maze,":[97],"RoFB":[98],"yields":[99],"clear":[100],"gains":[101],"in":[102,112],"four":[103,115],"of":[104,143],"six":[105],"model-task":[106],"pairs":[107,117],"small":[110],"gain":[111],"one.":[113],"The":[114],"positive":[116],"achieve":[118],"performance":[119],"unattainable":[120],"merely":[122],"selecting":[127],"from":[128],"multiple":[129],"comparable":[132],"lower":[134],"computational":[135],"cost.":[136],"These":[137],"results":[138],"suggest":[139],"closed-loop":[141],"steering":[142],"serve":[147],"complementary":[150],"control":[152],"mechanism":[153],"models.":[157]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-08-27T00:00:00"}