{"id":"https://openalex.org/W4416600577","doi":"https://doi.org/10.48550/arxiv.2505.22375","title":"Pangu Embedded: An Efficient Dual-system LLM Reasoner with Metacognition","display_name":"Pangu Embedded: An Efficient Dual-system LLM Reasoner with Metacognition","publication_year":2025,"publication_date":"2025-05-28","ids":{"openalex":"https://openalex.org/W4416600577","doi":"https://doi.org/10.48550/arxiv.2505.22375"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2505.22375","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2505.22375","pdf_url":"https://arxiv.org/pdf/2505.22375","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":null,"license_id":null,"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/2505.22375","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Chen, Hanting","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Hanting","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5072845284","display_name":"Yasheng Wang","orcid":"https://orcid.org/0000-0002-3221-0470","observed_orcids":["https://orcid.org/0000-0002-3221-0470"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yasheng","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Han, Kai","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Han, Kai","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100688935","display_name":"Li Dong","orcid":"https://orcid.org/0000-0002-1612-4716","observed_orcids":["https://orcid.org/0000-0002-1612-4716"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Dong","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100412858","display_name":"Lin Li","orcid":"https://orcid.org/0000-0002-2516-5827","observed_orcids":["https://orcid.org/0000-0002-2516-5827"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Lin","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Bi, Zhenni","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bi, Zhenni","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100782329","display_name":"Jinpeng Li","orcid":"https://orcid.org/0009-0007-2572-117X","observed_orcids":["https://orcid.org/0009-0007-2572-117X","https://orcid.org/0000-0001-8701-2642"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Jinpeng","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5115695530","display_name":"Haoyu Wang","orcid":"https://orcid.org/0000-0003-1100-8633","observed_orcids":["https://orcid.org/0000-0003-1100-8633","https://orcid.org/0009-0004-3221-6727","https://orcid.org/0009-0005-5788-2926","https://orcid.org/0009-0005-8191-2099"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Haoyu","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5012014905","display_name":"Fei Mi","orcid":"https://orcid.org/0000-0001-6358-9922","observed_orcids":["https://orcid.org/0000-0001-6358-9922"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mi, Fei","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5018719202","display_name":"Mingjian Zhu","orcid":"https://orcid.org/0000-0003-1726-9148","observed_orcids":["https://orcid.org/0000-0003-1726-9148"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhu, Mingjian","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100372395","display_name":"Bin Wang","orcid":"https://orcid.org/0000-0002-8755-6401","observed_orcids":["https://orcid.org/0000-0002-8755-6401"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Bin","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Song, Kaikai","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Song, Kaikai","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5077057825","display_name":"Yifei Fu","orcid":"https://orcid.org/0000-0001-9325-5114","observed_orcids":["https://orcid.org/0000-0001-9325-5114"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fu, Yifei","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101346649","display_name":"Xu He","orcid":"https://orcid.org/0009-0001-0419-5146","observed_orcids":["https://orcid.org/0009-0001-0419-5146"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"He, Xu","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101248003","display_name":"Yu Luo","orcid":"https://orcid.org/0009-0008-2652-6138","observed_orcids":["https://orcid.org/0009-0008-2652-6138"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Luo, Yu","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101180259","display_name":"Chong Zhu","orcid":"https://orcid.org/0000-0002-6434-6006","observed_orcids":["https://orcid.org/0000-0002-6434-6006"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhu, Chong","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5012644724","display_name":"Quan Sophia He","orcid":"https://orcid.org/0000-0002-0844-252X","observed_orcids":["https://orcid.org/0000-0002-0844-252X"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"He, Quan","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101077003","display_name":"Xueyu Wu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Xueyu","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5070006732","display_name":"Wei He","orcid":"https://orcid.org/0000-0002-3671-8650","observed_orcids":["https://orcid.org/0000-0002-3671-8650"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"He, Wei","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103144672","display_name":"Hailin Hu","orcid":"https://orcid.org/0000-0003-3800-1511","observed_orcids":["https://orcid.org/0000-0003-3800-1511"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hu, Hailin","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5111186638","display_name":"Yehui Tang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tang, Yehui","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5074103823","display_name":"Dacheng Tao","orcid":"https://orcid.org/0000-0001-7225-5449","observed_orcids":["https://orcid.org/0000-0001-7225-5449","https://orcid.org/0000-0002-5979-578X","https://orcid.org/0000-0002-9209-036X"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tao, Dacheng","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006817088","display_name":"Xinghao Chen","orcid":"https://orcid.org/0000-0002-2102-8235","observed_orcids":["https://orcid.org/0000-0002-2102-8235"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Xinghao","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100727355","display_name":"Yunhe Wang","orcid":"https://orcid.org/0000-0002-0013-4530","observed_orcids":["https://orcid.org/0000-0002-0013-4530","https://orcid.org/0000-0003-3643-6503"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yunhe","raw_affiliation_strings":["and Other Contributors"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"and Other Contributors","institution_ids":[]}]}],"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":0.8036999702453613,"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":0.8036999702453613,"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":0.07199999690055847,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.03750000149011612,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/llm-reasoning","display_name":"LLM reasoning","score":0.7639999985694885},{"id":"https://openalex.org/keywords/metacognition","display_name":"metacognition","score":0.9160000085830688},{"id":"https://openalex.org/keywords/fast-and-slow-modes","display_name":"fast and slow modes","score":0.6340000033378601},{"id":"https://openalex.org/keywords/knowledge-distillation","display_name":"knowledge distillation","score":0.6320000290870667},{"id":"https://openalex.org/keywords/mode-choice","display_name":"mode choice","score":0.3610000014305115},{"id":"https://openalex.org/keywords/mathematical-reasoning","display_name":"mathematical reasoning","score":0.5329999923706055},{"id":"https://openalex.org/keywords/code-reasoning","display_name":"code reasoning","score":0.5070000290870667},{"id":"https://openalex.org/keywords/question-answering","display_name":"question answering","score":0.5799999833106995}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.880299985408783},{"id":"https://openalex.org/C9616225","wikidata":"https://www.wikidata.org/wiki/Q3929429","display_name":"Semantic reasoner","level":2,"score":0.6409000158309937},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.6272000074386597},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.5404999852180481},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5134000182151794},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.4794999957084656},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4754999876022339},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3921999931335449},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.37220001220703125},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.35350000858306885},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.33410000801086426},{"id":"https://openalex.org/C48677424","wikidata":"https://www.wikidata.org/wiki/Q6888088","display_name":"Mode (computer interface)","level":2,"score":0.3255999982357025},{"id":"https://openalex.org/C143587482","wikidata":"https://www.wikidata.org/wiki/Q1543216","display_name":"Iterative and incremental development","level":2,"score":0.32170000672340393},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.3172999918460846},{"id":"https://openalex.org/C159032336","wikidata":"https://www.wikidata.org/wiki/Q2488768","display_name":"Non-monotonic logic","level":2,"score":0.30300000309944153},{"id":"https://openalex.org/C106189395","wikidata":"https://www.wikidata.org/wiki/Q176789","display_name":"Markov decision process","level":3,"score":0.3025999963283539},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.29750001430511475},{"id":"https://openalex.org/C12713177","wikidata":"https://www.wikidata.org/wiki/Q1900281","display_name":"Perspective (graphical)","level":2,"score":0.2946000099182129},{"id":"https://openalex.org/C174998907","wikidata":"https://www.wikidata.org/wiki/Q357662","display_name":"Work in process","level":2,"score":0.29010000824928284},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.2777999937534332},{"id":"https://openalex.org/C115901376","wikidata":"https://www.wikidata.org/wiki/Q184199","display_name":"Automation","level":2,"score":0.2635999917984009},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.2630999982357025},{"id":"https://openalex.org/C166088908","wikidata":"https://www.wikidata.org/wiki/Q308495","display_name":"Abductive reasoning","level":2,"score":0.25999999046325684},{"id":"https://openalex.org/C203479927","wikidata":"https://www.wikidata.org/wiki/Q5165939","display_name":"Controller (irrigation)","level":2,"score":0.25769999623298645},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.2565999925136566}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2505.22375","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2505.22375","pdf_url":"https://arxiv.org/pdf/2505.22375","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2505.22375","doi":"https://doi.org/10.48550/arxiv.2505.22375","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2505.22375","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2505.22375","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2505.22375","pdf_url":"https://arxiv.org/pdf/2505.22375","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"sustainable_development_goals_aurora":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7114801189","https://openalex.org/W6929732311","https://openalex.org/W4415336492","https://openalex.org/W4417100137","https://openalex.org/W7158423115","https://openalex.org/W4416172925","https://openalex.org/W4415021946","https://openalex.org/W7138195392","https://openalex.org/W7134071283","https://openalex.org/W7213401047"],"abstract_inverted_index":{"This":[0,72,151],"work":[1],"presents":[2],"Pangu":[3,26,134,191],"Embedded,":[4],"an":[5,59,162],"efficient":[6],"Large":[7],"Language":[8],"Model":[9],"(LLM)":[10],"reasoner":[11],"developed":[12],"on":[13,78,181],"Ascend":[14,79],"Neural":[15],"Processing":[16],"Units":[17],"(NPUs),":[18],"featuring":[19],"flexible":[20],"fast":[21],"and":[22,33,116,123,143,161,176,187,201,207],"slow":[23],"thinking":[24],"capabilities.":[25],"Embedded":[27,135,192],"addresses":[28],"the":[29,54],"significant":[30],"computational":[31,171],"costs":[32],"inference":[34],"latency":[35,175],"challenges":[36],"prevalent":[37],"in":[38],"existing":[39],"reasoning-optimized":[40],"LLMs.":[41],"We":[42],"propose":[43],"a":[44,83,101,130,137,144,212,218],"two-stage":[45],"training":[46],"framework":[47,152],"for":[48,120,140,148,158,221],"its":[49],"construction.":[50],"In":[51],"Stage":[52,127],"1,":[53],"model":[55,65,215],"is":[56,73,98],"finetuned":[57],"via":[58],"iterative":[60],"distillation":[61],"process,":[62],"incorporating":[63],"inter-iteration":[64],"merging":[66],"to":[67,173],"effectively":[68],"aggregate":[69],"complementary":[70],"knowledge.":[71],"followed":[74],"by":[75,82,100],"reinforcement":[76],"learning":[77],"clusters,":[80],"optimized":[81],"latency-tolerant":[84],"scheduler":[85],"that":[86,168,190],"combines":[87],"stale":[88],"synchronous":[89],"parallelism":[90],"with":[91,136,193],"prioritized":[92],"data":[93],"queues.":[94],"The":[95],"RL":[96],"process":[97],"guided":[99],"Multi-source":[102],"Adaptive":[103],"Reward":[104],"System":[105],"(MARS),":[106],"which":[107],"generates":[108],"dynamic,":[109],"task-specific":[110],"reward":[111],"signals":[112],"using":[113],"deterministic":[114],"metrics":[115],"lightweight":[117],"LLM":[118,227],"evaluators":[119],"mathematics,":[121],"coding,":[122],"general":[124],"problem-solving":[125],"tasks.":[126],"2":[128],"introduces":[129],"dual-system":[131],"framework,":[132],"endowing":[133],"\"fast\"":[138],"mode":[139,147,156,165],"routine":[141],"queries":[142],"deeper":[145],"\"slow\"":[146],"complex":[149],"inference.":[150],"offers":[153],"both":[154],"manual":[155],"switching":[157],"user":[159],"control":[160],"automatic,":[163],"complexity-aware":[164],"selection":[166],"mechanism":[167],"dynamically":[169],"allocates":[170],"resources":[172],"balance":[174],"reasoning":[177,209],"depth.":[178],"Experimental":[179],"results":[180],"benchmarks":[182],"including":[183],"AIME":[184],"2024,":[185],"GPQA,":[186],"LiveCodeBench":[188],"demonstrate":[189],"7B":[194],"parameters,":[195],"outperforms":[196],"similar-size":[197],"models":[198],"like":[199],"Qwen3-8B":[200],"GLM4-9B.":[202],"It":[203],"delivers":[204],"rapid":[205],"responses":[206],"state-of-the-art":[208],"quality":[210],"within":[211],"single,":[213],"unified":[214],"architecture,":[216],"highlighting":[217],"promising":[219],"direction":[220],"developing":[222],"powerful":[223],"yet":[224],"practically":[225],"deployable":[226],"reasoners.":[228]},"counts_by_year":[],"updated_date":"2026-10-16T09:59:22.235173","created_date":"2025-10-10T00:00:00"}