{"id":"https://openalex.org/W7161830278","doi":"https://doi.org/10.48550/arxiv.2605.18851","title":"STRIDE: Learnable Stepwise Language Feedback for LLM Reasoning","display_name":"STRIDE: Learnable Stepwise Language Feedback for LLM Reasoning","publication_year":2026,"publication_date":"2026-05-13","ids":{"openalex":"https://openalex.org/W7161830278","doi":"https://doi.org/10.48550/arxiv.2605.18851"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.18851","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.18851","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":"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.18851","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5136544066","display_name":"Junjie Zhang","orcid":"https://orcid.org/0000-0002-6826-1798"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Junjie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136598261","display_name":"Guozheng Ma","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ma, Guozheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5090854835","display_name":"Shunyu Liu","orcid":"https://orcid.org/0000-0003-0584-9129"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Shunyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136594687","display_name":"Zetian Hu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hu, Zetian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136534314","display_name":"Yongcheng Jing","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jing, Yongcheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069007219","display_name":"Ting-En Lin","orcid":"https://orcid.org/0000-0003-0650-7286"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lin, Ting-En","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136582197","display_name":"Yongbin Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Yongbin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5136562196","display_name":"Dacheng Tao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tao, Dacheng","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/T10028","display_name":"Topic Modeling","score":0.42419999837875366,"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":0.42419999837875366,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.10379999876022339,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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":0.09220000356435776,"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","display_name":"Reinforcement learning","score":0.5907999873161316},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.5230000019073486},{"id":"https://openalex.org/keywords/generator","display_name":"Generator (circuit theory)","score":0.44119998812675476},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.4262999892234802},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.42080000042915344},{"id":"https://openalex.org/keywords/trajectory","display_name":"Trajectory","score":0.3330000042915344}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7339000105857849},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5907999873161316},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.5230000019073486},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4867999851703644},{"id":"https://openalex.org/C2780992000","wikidata":"https://www.wikidata.org/wiki/Q17016113","display_name":"Generator (circuit theory)","level":3,"score":0.44119998812675476},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.4262999892234802},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.42080000042915344},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.3330000042915344},{"id":"https://openalex.org/C2777212361","wikidata":"https://www.wikidata.org/wiki/Q5127848","display_name":"Class (philosophy)","level":2,"score":0.3167000114917755},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.3109000027179718},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2865000069141388},{"id":"https://openalex.org/C2776257435","wikidata":"https://www.wikidata.org/wiki/Q1576430","display_name":"Bandwidth (computing)","level":2,"score":0.2705000042915344},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2637999951839447},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.26010000705718994},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.2549000084400177}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.18851","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.18851","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.18851","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.18851","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":"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"},"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":{"Recent":[0],"advances":[1],"in":[2,189],"Reinforcement":[3],"Learning":[4],"(RL)":[5],"have":[6],"underscored":[7],"its":[8],"potential":[9],"for":[10,66,201],"incentivizing":[11],"reasoning":[12,139,165],"capabilities":[13],"of":[14,196],"Large":[15],"Language":[16],"Models":[17],"(LLMs).":[18],"However,":[19],"existing":[20],"step-level":[21],"efforts":[22],"suffer":[23],"from":[24,90],"costly":[25],"annotations":[26],"that":[27,86,168],"limit":[28],"domain":[29],"coverage,":[30],"while":[31,114],"scalar":[32,91,183],"scores":[33],"further":[34],"impose":[35],"an":[36],"information":[37],"bottleneck,":[38],"offering":[39],"insufficient":[40],"semantic":[41],"bandwidth":[42],"to":[43,93,137],"improve":[44],"intermediate":[45,142],"decisions.":[46,146],"Alternative":[47],"language-critique":[48],"approaches,":[49],"which":[50],"rely":[51],"on":[52,163,179],"frozen":[53],"or":[54,158],"external":[55,112],"critics,":[56],"provide":[57],"richer":[58],"textual":[59],"feedback":[60,200],"but":[61],"lack":[62],"the":[63,135,194],"scalability":[64],"needed":[65],"sustained":[67,116],"policy":[68,117,153],"improvement.":[69],"In":[70],"this":[71],"work,":[72],"we":[73,99],"propose":[74],"language-driven":[75],"stepwise":[76,95,126,198],"trajectory":[77,148],"redirection,":[78],"termed":[79],"as":[80,174,176],"STRIDE,":[81],"a":[82,101,104],"novel":[83],"training":[84],"framework":[85],"shifts":[87],"process":[88],"supervision":[89],"rewards":[92],"learnable":[94,197],"language":[96,127,199],"feedback.":[97,161],"Specifically,":[98],"co-train":[100],"generator":[102,136],"and":[103,131],"generative":[105],"verifier":[106,122,160],"using":[107],"only":[108],"outcome-based":[109],"rewards,":[110],"eliminating":[111],"annotations,":[113],"delivering":[115],"improvement":[118],"through":[119],"jointly":[120],"aligned":[121],"training.":[123],"The":[124,147],"verifier's":[125],"critiques":[128],"explicitly":[129],"localize":[130],"explain":[132],"failures,":[133],"enabling":[134],"redirect":[138],"trajectories":[140],"at":[141],"steps":[143],"toward":[144],"alternative":[145],"redirection":[149],"design":[150],"guarantees":[151],"harmless":[152],"improvement,":[154],"even":[155],"under":[156],"noisy":[157],"suboptimal":[159],"Experiments":[162],"diverse":[164],"benchmarks":[166],"show":[167],"STRIDE":[169],"significantly":[170],"outperforms":[171],"state-of-the-art":[172],"baselines,":[173],"well":[175],"achieving":[177],"breakthroughs":[178],"zero-pass-rate":[180],"problems":[181],"where":[182],"methods":[184],"yield":[185],"no":[186],"learning":[187],"signal":[188],"our":[190],"ablation":[191],"studies,":[192],"demonstrating":[193],"effectiveness":[195],"enhancing":[202],"LLM":[203],"reasoning.":[204]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-21T00:00:00"}