{"id":"https://openalex.org/W7154968036","doi":"https://doi.org/10.48550/arxiv.2604.16029","title":"Cut Your Losses! Learning to Prune Paths Early for Efficient Parallel Reasoning","display_name":"Cut Your Losses! Learning to Prune Paths Early for Efficient Parallel Reasoning","publication_year":2026,"publication_date":"2026-04-17","ids":{"openalex":"https://openalex.org/W7154968036","doi":"https://doi.org/10.48550/arxiv.2604.16029"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.16029","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.16029","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2604.16029","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5134033200","display_name":"Jiaxi Bi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bi, Jiaxi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134006965","display_name":"Tongxu Luo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Luo, Tongxu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134063289","display_name":"Wenyu Du","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Du, Wenyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134068157","display_name":"Zhengyang Tang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tang, Zhengyang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5134012405","display_name":"Benyou Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Benyou","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":5,"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":0.18369999527931213,"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":0.18369999527931213,"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/T10906","display_name":"AI-based Problem Solving and Planning","score":0.10350000113248825,"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/T11273","display_name":"Advanced Graph Neural Networks","score":0.1005999967455864,"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/boosting","display_name":"Boosting (machine learning)","score":0.6826000213623047},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.6504999995231628},{"id":"https://openalex.org/keywords/learnability","display_name":"Learnability","score":0.6086000204086304},{"id":"https://openalex.org/keywords/path","display_name":"Path (computing)","score":0.5462999939918518},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.5012999773025513},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.4255000054836273},{"id":"https://openalex.org/keywords/correctness","display_name":"Correctness","score":0.39719998836517334}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7537000179290771},{"id":"https://openalex.org/C46686674","wikidata":"https://www.wikidata.org/wiki/Q466303","display_name":"Boosting (machine learning)","level":2,"score":0.6826000213623047},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.6504999995231628},{"id":"https://openalex.org/C2777723229","wikidata":"https://www.wikidata.org/wiki/Q4367921","display_name":"Learnability","level":2,"score":0.6086000204086304},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5462999939918518},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.5462999939918518},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.5012999773025513},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4580000042915344},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.4255000054836273},{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.39719998836517334},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.3537999987602234},{"id":"https://openalex.org/C141603448","wikidata":"https://www.wikidata.org/wiki/Q134830","display_name":"Prefix","level":2,"score":0.3407999873161316},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.31540000438690186},{"id":"https://openalex.org/C2780069185","wikidata":"https://www.wikidata.org/wiki/Q7977945","display_name":"Equivalence (formal languages)","level":2,"score":0.3100000023841858},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3018999993801117},{"id":"https://openalex.org/C120936955","wikidata":"https://www.wikidata.org/wiki/Q2155640","display_name":"Empirical research","level":2,"score":0.2824000120162964},{"id":"https://openalex.org/C115051666","wikidata":"https://www.wikidata.org/wiki/Q6522493","display_name":"Ranging","level":2,"score":0.2556999921798706},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.2533999979496002},{"id":"https://openalex.org/C58642233","wikidata":"https://www.wikidata.org/wiki/Q8269924","display_name":"Taxonomy (biology)","level":2,"score":0.2531000077724457}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.16029","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.16029","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2604.16029","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.16029","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"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":{"Parallel":[0],"reasoning":[1],"enhances":[2],"Large":[3],"Reasoning":[4],"Models":[5],"(LRMs)":[6],"but":[7],"incurs":[8],"prohibitive":[9],"costs":[10],"due":[11],"to":[12,91,103,128,144],"futile":[13],"paths":[14],"caused":[15],"by":[16,53],"early":[17],"errors.":[18],"To":[19],"mitigate":[20],"this,":[21],"path":[22,49],"pruning":[23],"at":[24,155],"the":[25,44,68,110],"prefix":[26],"level":[27],"is":[28],"essential,":[29],"yet":[30],"existing":[31,104],"research":[32],"remains":[33],"fragmented":[34],"without":[35],"a":[36],"standardized":[37],"framework.":[38],"In":[39],"this":[40],"work,":[41],"we":[42,107,136],"propose":[43],"first":[45],"systematic":[46],"taxonomy":[47],"of":[48,71,78,112],"pruning,":[50],"categorizing":[51],"methods":[52],"their":[54],"signal":[55],"source":[56],"(internal":[57],"vs.":[58,63],"external)":[59],"and":[60,100,151],"learnability":[61],"(learnable":[62],"non-learnable).":[64],"This":[65],"classification":[66],"reveals":[67],"unexplored":[69],"potential":[70],"learnable":[72],"internal":[73],"methods,":[74],"motivating":[75],"our":[76,138],"proposal":[77],"STOP":[79,96,113],"(Super":[80],"TOken":[81],"for":[82,119],"Pruning).":[83],"Extensive":[84],"evaluations":[85],"across":[86],"LRMs":[87],"ranging":[88],"from":[89,126],"1.5B":[90],"20B":[92],"parameters":[93],"demonstrate":[94],"that":[95],"achieves":[97],"superior":[98],"effectiveness":[99],"efficiency":[101],"compared":[102],"baselines.":[105],"Furthermore,":[106],"rigorously":[108],"validate":[109],"scalability":[111],"under":[114,131],"varying":[115],"compute":[116,133],"budgets":[117],"-":[118],"instance,":[120],"boosting":[121],"GPT-OSS-20B":[122],"accuracy":[123],"on":[124],"AIME25":[125],"84%":[127],"nearly":[129],"90%":[130],"fixed":[132],"budgets.":[134],"Finally,":[135],"distill":[137],"findings":[139],"into":[140],"formalized":[141],"empirical":[142],"guidelines":[143],"facilitate":[145],"optimal":[146],"real-world":[147],"deployment.":[148],"Code,":[149],"data":[150],"models":[152],"are":[153],"available":[154],"https://bijiaxihh.github.io/STOP":[156]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-04-21T00:00:00"}