{"id":"https://openalex.org/W7225626913","doi":"https://doi.org/10.21203/rs.3.rs-11669746/v1","title":"Adaptive Inference-Time Reasoning for Large Language Models via Cross-Task Sparse Policy Transfer","display_name":"Adaptive Inference-Time Reasoning for Large Language Models via Cross-Task Sparse Policy Transfer","publication_year":2026,"publication_date":"2026-11-10","ids":{"openalex":"https://openalex.org/W7225626913","doi":"https://doi.org/10.21203/rs.3.rs-11669746/v1"},"language":"en","primary_location":{"id":"doi:10.21203/rs.3.rs-11669746/v1","doi":"https://doi.org/10.21203/rs.3.rs-11669746/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-11669746/v1","pdf_url":"https://www.researchsquare.com/article/rs-11669746/latest.pdf","source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Research Square","raw_type":"posted-content"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://www.researchsquare.com/article/rs-11669746/latest.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5161351591","display_name":"Duo Xu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I12912129","display_name":"Northeastern University","ror":"https://ror.org/04t5xt781","country_code":"US","type":"education","lineage":["https://openalex.org/I12912129"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Duo Xu","raw_affiliation_strings":["Northeastern University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Northeastern University","institution_ids":["https://openalex.org/I12912129"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5157515918","display_name":"Zitan Chen","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I74973139","display_name":"Carnegie Mellon University","ror":"https://ror.org/05x2bcf33","country_code":"US","type":"education","lineage":["https://openalex.org/I74973139"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zitan Chen","raw_affiliation_strings":["Carnegie Mellon University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Carnegie Mellon University","institution_ids":["https://openalex.org/I74973139"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5103041351","display_name":"Chenyi Huang","orcid":"https://orcid.org/0000-0002-7664-0720","observed_orcids":["https://orcid.org/0000-0002-7664-0720"]},"institutions":[{"id":"https://openalex.org/I145311948","display_name":"Johns Hopkins University","ror":"https://ror.org/00za53h95","country_code":"US","type":"education","lineage":["https://openalex.org/I145311948"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Chenyi Huang","raw_affiliation_strings":["Johns Hopkins University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Johns Hopkins University","institution_ids":["https://openalex.org/I145311948"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":[],"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/T10028","display_name":"Topic Modeling","score":0.6904000043869019,"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":0.6904000043869019,"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":0.16940000653266907,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.022700000554323196,"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/large-language-models","display_name":"large language models","score":0.7250000238418579},{"id":"https://openalex.org/keywords/qwen2-5","display_name":"Qwen2.5","score":0.7369999885559082},{"id":"https://openalex.org/keywords/prompt-injection","display_name":"prompt injection","score":0.4050000011920929},{"id":"https://openalex.org/keywords/rast","display_name":"radioallergosorbent test","score":0.5090000033378601}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7250000238418579},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.696399986743927},{"id":"https://openalex.org/C199521495","wikidata":"https://www.wikidata.org/wiki/Q181487","display_name":"Audit","level":2,"score":0.5569999814033508},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5473999977111816},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5117999911308289},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.46389999985694885},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.4397999942302704},{"id":"https://openalex.org/C2776175482","wikidata":"https://www.wikidata.org/wiki/Q1195816","display_name":"Transfer (computing)","level":2,"score":0.4336000084877014},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.42179998755455017},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.41040000319480896},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3779999911785126},{"id":"https://openalex.org/C134121241","wikidata":"https://www.wikidata.org/wiki/Q899301","display_name":"Yield (engineering)","level":2,"score":0.31049999594688416},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.29339998960494995},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2736000120639801},{"id":"https://openalex.org/C165838908","wikidata":"https://www.wikidata.org/wiki/Q736777","display_name":"Calibration","level":2,"score":0.2660999894142151},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.26350000500679016},{"id":"https://openalex.org/C2983203078","wikidata":"https://www.wikidata.org/wiki/Q255166","display_name":"Information gain","level":2,"score":0.2630999982357025},{"id":"https://openalex.org/C2775973920","wikidata":"https://www.wikidata.org/wiki/Q3252726","display_name":"Selection algorithm","level":3,"score":0.2567000091075897}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.21203/rs.3.rs-11669746/v1","doi":"https://doi.org/10.21203/rs.3.rs-11669746/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-11669746/v1","pdf_url":"https://www.researchsquare.com/article/rs-11669746/latest.pdf","source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Research Square","raw_type":"posted-content"}],"best_oa_location":{"id":"doi:10.21203/rs.3.rs-11669746/v1","doi":"https://doi.org/10.21203/rs.3.rs-11669746/v1","is_oa":true,"landing_page_url":"https://doi.org/10.21203/rs.3.rs-11669746/v1","pdf_url":"https://www.researchsquare.com/article/rs-11669746/latest.pdf","source":{"id":"https://openalex.org/S4306525896","display_name":"Research Square","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"Research Square","raw_type":"posted-content"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7225626913.pdf","grobid_xml":"https://content.openalex.org/works/W7225626913.grobid-xml"},"referenced_works_count":8,"referenced_works":["https://openalex.org/W2110065044","https://openalex.org/W3195140297","https://openalex.org/W3210904070","https://openalex.org/W4385571157","https://openalex.org/W4389520688","https://openalex.org/W4414898977","https://openalex.org/W4417536778","https://openalex.org/W7133224126"],"related_works":["https://openalex.org/W4402671856","https://openalex.org/W7203697700","https://openalex.org/W7169107258","https://openalex.org/W7166824454","https://openalex.org/W7130927144","https://openalex.org/W7131793712","https://openalex.org/W7203911768","https://openalex.org/W7223122654","https://openalex.org/W7104972725","https://openalex.org/W7211939736"],"abstract_inverted_index":{"Abstract":[0],"Sparse":[1],"guide":[2,84,104],"queries":[3],"need":[4],"not":[5],"yield":[6],"proportional":[7],"computation":[8],"savings":[9],"or":[10],"better":[11],"reasoning.":[12],"We":[13],"audit":[14,96],"a":[15,34,65],"frozen":[16],"Qwen2.5-0.5B":[17],"base/RL":[18],"pair":[19],"on":[20,37],"1,012":[21],"reference":[22],"questions":[23,40],"and":[24,53,102],"an":[25],"exploratory":[26],"128-question":[27],"generation":[28,70],"pilot.":[29],"Aligned":[30],"policy":[31],"differences":[32],"outperform":[33],"token-permuted":[35],"control":[36],"192":[38],"new":[39],"per":[41],"math/code":[42],"domain.":[43],"Local-harm":[44],"AUROC":[45],"reaches":[46],"0.826/0.871,":[47],"but":[48,83],"calibration,":[49],"score":[50],"ranking,":[51],"ablations,":[52],"threshold":[54],"scans":[55],"expose":[56],"selection":[57,100],"failures.":[58],"Adaptive":[59],"gains":[60],"of":[61],"0.00011/0.00039":[62],"nat/token":[63],"lack":[64],"Benjamini\u2013Hochberg-adjusted":[66],"advantage":[67],"over":[68],"base;":[69],"accuracy":[71],"is":[72],"37.5%":[73],"versus":[74],"38.3%.":[75],"Against":[76],"full-token":[77],"RAST-form":[78],"injection,":[79],"calls":[80],"fall":[81],"83.6%":[82],"input":[85],"tokens":[86],"only":[87],"25.5%,":[88],"chiefly":[89],"through":[90],"avoided":[91],"prompt":[92],"prefill.":[93],"This":[94],"single-checkpoint":[95],"distinguishes":[97],"signal":[98],"information,":[99],"quality,":[101],"actual":[103],"work.":[105]},"counts_by_year":[],"updated_date":"2026-11-11T12:16:26.678958","created_date":"2026-11-10T00:00:00"}