{"id":"https://openalex.org/W7223875253","doi":"https://doi.org/10.48550/arxiv.2610.35226","title":"Trainable Boundary-Keeping Where No Formal Answer Exists: A Controlled Knock-In Study of Reward Training in Evidence-Sufficiency Judgment","display_name":"Trainable Boundary-Keeping Where No Formal Answer Exists: A Controlled Knock-In Study of Reward Training in Evidence-Sufficiency Judgment","publication_year":2026,"publication_date":"2026-10-26","ids":{"openalex":"https://openalex.org/W7223875253","doi":"https://doi.org/10.48550/arxiv.2610.35226"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2610.35226","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.35226","pdf_url":"https://arxiv.org/pdf/2610.35226","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/2610.35226","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5147664576","display_name":"Yajie Yin","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Yin, YaJie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":["https://openalex.org/A5147664576"],"corresponding_institution_ids":[],"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/T11010","display_name":"Logic, Reasoning, and Knowledge","score":0.15919999778270721,"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/T11010","display_name":"Logic, Reasoning, and Knowledge","score":0.15919999778270721,"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":0.09969999641180038,"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/T10315","display_name":"Decision-Making and Behavioral Economics","score":0.05590000003576279,"subfield":{"id":"https://openalex.org/subfields/1800","display_name":"General Decision Sciences"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/automated-verification","display_name":"automated verification","score":0.3109999895095825},{"id":"https://openalex.org/keywords/causal-claims","display_name":"causal claims","score":0.5770000219345093},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.382999986410141}],"concepts":[{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.7419999837875366},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.5324000120162964},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.522599995136261},{"id":"https://openalex.org/C204434341","wikidata":"https://www.wikidata.org/wiki/Q357789","display_name":"Adjudication","level":2,"score":0.48829999566078186},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.45879998803138733},{"id":"https://openalex.org/C94624232","wikidata":"https://www.wikidata.org/wiki/Q3150667","display_name":"Indeterminate","level":2,"score":0.4392000138759613},{"id":"https://openalex.org/C153269930","wikidata":"https://www.wikidata.org/wiki/Q430001","display_name":"Decidability","level":2,"score":0.40529999136924744},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.39629998803138733},{"id":"https://openalex.org/C95623464","wikidata":"https://www.wikidata.org/wiki/Q1096149","display_name":"Classifier (UML)","level":2,"score":0.3495999872684479},{"id":"https://openalex.org/C77805123","wikidata":"https://www.wikidata.org/wiki/Q161272","display_name":"Social psychology","level":1,"score":0.34290000796318054},{"id":"https://openalex.org/C77553402","wikidata":"https://www.wikidata.org/wiki/Q13222579","display_name":"Upper and lower bounds","level":2,"score":0.3391999900341034},{"id":"https://openalex.org/C2777267654","wikidata":"https://www.wikidata.org/wiki/Q3519023","display_name":"Test (biology)","level":2,"score":0.33899998664855957},{"id":"https://openalex.org/C2777530160","wikidata":"https://www.wikidata.org/wiki/Q41796","display_name":"Sentence","level":2,"score":0.32440000772476196},{"id":"https://openalex.org/C102931765","wikidata":"https://www.wikidata.org/wiki/Q189673","display_name":"Dissociation (chemistry)","level":2,"score":0.32339999079704285},{"id":"https://openalex.org/C136979486","wikidata":"https://www.wikidata.org/wiki/Q773483","display_name":"Existential quantification","level":2,"score":0.32260000705718994},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3221000134944916},{"id":"https://openalex.org/C151319957","wikidata":"https://www.wikidata.org/wiki/Q752739","display_name":"Asynchronous communication","level":2,"score":0.28780001401901245},{"id":"https://openalex.org/C110251889","wikidata":"https://www.wikidata.org/wiki/Q1569697","display_name":"Model checking","level":2,"score":0.2840000092983246},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.2711000144481659},{"id":"https://openalex.org/C159718280","wikidata":"https://www.wikidata.org/wiki/Q5526353","display_name":"Gas meter prover","level":3,"score":0.26899999380111694},{"id":"https://openalex.org/C2778067643","wikidata":"https://www.wikidata.org/wiki/Q166507","display_name":"Interval (graph theory)","level":2,"score":0.26809999346733093},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.26350000500679016},{"id":"https://openalex.org/C134752490","wikidata":"https://www.wikidata.org/wiki/Q374182","display_name":"Logical consequence","level":2,"score":0.26179999113082886},{"id":"https://openalex.org/C2778012447","wikidata":"https://www.wikidata.org/wiki/Q1034415","display_name":"Scope (computer science)","level":2,"score":0.2533999979496002},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.2533000111579895},{"id":"https://openalex.org/C188198153","wikidata":"https://www.wikidata.org/wiki/Q1613840","display_name":"Limiting","level":2,"score":0.25099998712539673}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2610.35226","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.35226","pdf_url":"https://arxiv.org/pdf/2610.35226","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.2610.35226","doi":"https://doi.org/10.48550/arxiv.2610.35226","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2610.35226","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:2610.35226","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.35226","pdf_url":"https://arxiv.org/pdf/2610.35226","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":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7223875253.pdf","grobid_xml":"https://content.openalex.org/works/W7223875253.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7166814490","https://openalex.org/W7166808443","https://openalex.org/W7162848250","https://openalex.org/W7223140226","https://openalex.org/W4415276150","https://openalex.org/W7212123408","https://openalex.org/W7167284832","https://openalex.org/W7166800456","https://openalex.org/W7213927650","https://openalex.org/W7155098438"],"abstract_inverted_index":{"Reasoning-tuned":[0],"and":[1,63,156,169],"reward-trained":[2],"models":[3],"are":[4,79,247],"often":[5],"assumed":[6],"to":[7,87,103],"verify":[8],"more":[9],"reliably":[10],"because":[11],"they":[12],"reason":[13],"longer":[14],"or":[15,240],"receive":[16],"correctness":[17,39],"signals.":[18],"We":[19],"test":[20],"this":[21],"in":[22,173,195],"evidence-sufficiency":[23],"judgment:":[24],"deciding":[25],"whether":[26],"available":[27],"information":[28],"actually":[29],"supports":[30],"a":[31,34,42,50,60,64,157,228],"causal":[32],"claim,":[33],"domain":[35],"with":[36,56,72,115],"no":[37,73],"decidable":[38],"reward.":[40],"Using":[41],"pre-registered,":[43],"same-base,":[44],"controlled":[45,229],"knock-in":[46],"design,":[47],"we":[48,233],"compare":[49],"boundary-keeping":[51,183],"recipe":[52],"(SFT":[53],"then":[54],"GRPO)":[55],"its":[57],"own":[58],"baseline,":[59],"scrambled-pairing":[61],"placebo,":[62],"general-purpose":[65],"reasoning":[66],"model":[67,248],"under":[68],"an":[69,187],"action-pressure":[70],"protocol":[71],"evaluation":[74],"invitation.":[75],"All":[76],"load-bearing":[77],"outcomes":[78],"rated":[80],"by":[81],"two":[82,174],"separate":[83],"model-rating":[84],"sessions":[85,246],"masked":[86],"arm":[88,96],"labels;":[89],"automatic":[90,143],"judges":[91],"only":[92],"trigger.":[93],"The":[94,135,176,225],"trained":[95],"raises":[97],"spontaneous":[98,127],"correct":[99],"refusal":[100,145],"from":[101],"25%":[102],"100%":[104],"(40":[105],"per":[106],"arm;":[107],"Wilson":[108],"95%":[109],"CI":[110],"lower":[111],"bound":[112,154],"91.2%,":[113],"non-overlapping":[114],"the":[116,122,179,196,205,244],"baseline":[117],"interval":[118],"[14.2%,":[119],"40.2%]),":[120],"while":[121,190],"scrambled":[123],"placebo":[124],"collapses":[125],"(5%":[126],"/":[128],"7.5%":[129],"invited":[130],"correct;":[131],"91.1%":[132],"false":[133,144,171],"refusal).":[134],"effect":[136],"is":[137,150,178,184,227],"boundary-specific":[138],"rather":[139],"than":[140],"blanket":[141],"refusal:":[142],"on":[146],"281":[147],"well-formed":[148],"claims":[149],"1.1%":[151],"(Wilson":[152],"upper":[153],"3.09%),":[155],"targeted":[158],"18-item":[159],"arm-(c)":[160],"recheck":[161],"(3":[162],"flagged":[163],"plus":[164],"15":[165],"accepted)":[166],"yields":[167],"0/18":[168],"1/18":[170],"refusals":[172],"sessions.":[175],"dissociation":[177],"main":[180],"result:":[181],"final":[182],"trainable":[185],"without":[186],"inference-time":[188],"verifier,":[189],"observed":[191],"active":[192],"numeric/calculation":[193],"checking":[194],"H3":[197],"assay":[198],"remains":[199],"near":[200],"zero;":[201],"SFT":[202],"saturation":[203],"makes":[204],"GRPO":[206],"increment":[207],"not":[208,223,235,250],"separately":[209],"identifiable.":[210],"A":[211],"pre-registered":[212],"mechanism":[213],"adjudication":[214],"favors":[215],"reward":[216],"internalization":[217],"of":[218],"final-action":[219],"status":[220],"(H2,":[221],"best-supported,":[222],"proven).":[224],"scope":[226],"14B":[230],"same-family":[231],"study:":[232],"do":[234],"claim":[236],"proof,":[237],"cross-family":[238],"generality,":[239],"independent":[241],"human-expert":[242],"validation;":[243],"rating":[245],"sessions,":[249],"human":[251],"reviewers.":[252]},"counts_by_year":[],"updated_date":"2026-10-30T12:39:25.321706","created_date":"2026-10-28T00:00:00"}