{"id":"https://openalex.org/W4416034123","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.1020","title":"Fast Quiet-STaR: Thinking Without Thought Tokens","display_name":"Fast Quiet-STaR: Thinking Without Thought Tokens","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4416034123","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.1020"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.1020","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.1020","pdf_url":"https://aclanthology.org/2025.findings-emnlp.1020.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2025","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.findings-emnlp.1020.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5038403509","display_name":"Wei Huang","orcid":"https://orcid.org/0000-0003-4069-7343"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wei Huang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102630950","display_name":"Yizhe Xiong","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yizhe Xiong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100348117","display_name":"Xin Ye","orcid":"https://orcid.org/0000-0002-8429-0900"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xin Ye","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102366228","display_name":"Zhijie Deng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhijie Deng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101409147","display_name":"Hui Chen","orcid":"https://orcid.org/0000-0002-8010-8796"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hui Chen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5077882811","display_name":"Zijia Lin","orcid":"https://orcid.org/0000-0002-1390-7424"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zijia Lin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5057732142","display_name":"Guiguang Ding","orcid":"https://orcid.org/0000-0003-0137-9975"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guiguang Ding","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":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"18771","last_page":"18781"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11574","display_name":"Artificial Intelligence in Games","score":0.023900000378489494,"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/T11574","display_name":"Artificial Intelligence in Games","score":0.023900000378489494,"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/T12101","display_name":"Advanced Bandit Algorithms Research","score":0.01810000091791153,"subfield":{"id":"https://openalex.org/subfields/1803","display_name":"Management Science and Operations Research"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T10682","display_name":"Quantum Computing Algorithms and Architecture","score":0.01600000075995922,"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/inference","display_name":"Inference","score":0.7671999931335449},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.7368999719619751},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.5824999809265137},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.5188000202178955},{"id":"https://openalex.org/keywords/opportunistic-reasoning","display_name":"Opportunistic reasoning","score":0.3896999955177307},{"id":"https://openalex.org/keywords/range","display_name":"Range (aeronautics)","score":0.3749000132083893},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.34450000524520874},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.3361999988555908}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7800999879837036},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.7671999931335449},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.7368999719619751},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.5824999809265137},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5633000135421753},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.5188000202178955},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4108999967575073},{"id":"https://openalex.org/C86827895","wikidata":"https://www.wikidata.org/wiki/Q7098582","display_name":"Opportunistic reasoning","level":4,"score":0.3896999955177307},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.3749000132083893},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3481000065803528},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.34450000524520874},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.3361999988555908},{"id":"https://openalex.org/C2779439875","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Natural language understanding","level":3,"score":0.33309999108314514},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.32190001010894775},{"id":"https://openalex.org/C37335422","wikidata":"https://www.wikidata.org/wiki/Q6888134","display_name":"Model-based reasoning","level":3,"score":0.31700000166893005},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.3142000138759613},{"id":"https://openalex.org/C89288958","wikidata":"https://www.wikidata.org/wiki/Q7301504","display_name":"Reasoning system","level":2,"score":0.30559998750686646},{"id":"https://openalex.org/C195344581","wikidata":"https://www.wikidata.org/wiki/Q2555318","display_name":"Automated reasoning","level":2,"score":0.296099990606308},{"id":"https://openalex.org/C2983448237","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Language understanding","level":2,"score":0.2799000144004822},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.27869999408721924},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.2556000053882599}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-emnlp.1020","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.1020","pdf_url":"https://aclanthology.org/2025.findings-emnlp.1020.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2025","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.1020","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.1020","pdf_url":"https://aclanthology.org/2025.findings-emnlp.1020.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2025","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G338250432","display_name":null,"funder_award_id":"62441235","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416034123.pdf","grobid_xml":"https://content.openalex.org/works/W4416034123.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs)":[3],"have":[4],"achieved":[5],"impressive":[6],"performance":[7],"across":[8],"a":[9,68,86],"range":[10],"of":[11,77,95,160,177],"natural":[12],"language":[13],"processing":[14],"tasks.However,":[15],"recent":[16],"advances":[17],"demonstrate":[18,151],"that":[19,73,90,152],"further":[20,109],"gains-particularly":[21],"in":[22,126,158],"complex":[23],"reasoning":[24,46,51,71,79,107],"tasks-require":[25],"more":[26,69,103],"than":[27],"merely":[28],"scaling":[29],"up":[30],"model":[31,100],"sizes":[32],"or":[33],"training":[34,88],"data.One":[35],"promising":[36],"direction":[37],"is":[38],"to":[39,42,101,113],"enable":[40],"models":[41],"\"think\"":[43],"during":[44,139],"the":[45,75,93,99,114,132,164,189],"process.Recently,":[47],"Quiet-STaR":[48,128,154,157,170],"significantly":[49],"improves":[50],"by":[52],"generating":[53],"token-level":[54,78],"thought":[55,96,136],"traces,":[56],"but":[57],"incurs":[58],"substantial":[59],"inference":[60,166,191],"overhead.In":[61],"this":[62,111],"work,":[63],"we":[64],"propose":[65],"Fast":[66,127,153,169],"Quiet-STaR,":[67],"efficient":[70],"framework":[72],"preserves":[74],"benefits":[76],"while":[80,187],"reducing":[81],"computational":[82],"cost.Our":[83],"method":[84],"introduces":[85],"curriculum-learning-based":[87],"strategy":[89],"gradually":[91],"reduces":[92],"number":[94],"tokens,":[97],"enabling":[98],"internalize":[102],"abstract":[104],"and":[105,148,182],"concise":[106],"processes.We":[108],"extend":[110],"approach":[112],"standard":[115],"Next":[116],"Token":[117],"Prediction":[118],"(NTP)":[119],"setting":[120],"through":[121],"reinforcement":[122],"learning-based":[123],"fine-tuning,":[124],"resulting":[125],"NTP,":[129],"which":[130],"eliminates":[131],"need":[133],"for":[134],"explicit":[135],"token":[137],"generation":[138],"inference.Experiments":[140],"on":[141,179,184],"four":[142],"benchmark":[143],"datasets":[144],"with":[145],"Mistral":[146,180],"7B":[147,150,181],"Qwen2.5":[149,185],"consistently":[155],"outperforms":[156],"terms":[159],"average":[161,174],"accuracy":[162,175],"under":[163],"same":[165,190],"time":[167],"budget.Notably,":[168],"NTP":[171],"achieves":[172],"an":[173],"improvement":[176],"9%":[178],"5.7%":[183],"7B,":[186],"maintaining":[188],"latency.Our":[192],"code":[193],"will":[194],"be":[195],"available":[196],"at":[197],"https:":[198]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-11-08T00:00:00"}