{"id":"https://openalex.org/W7034070351","doi":null,"title":"Skeleton-of-thought: Large language models can do parallel decoding","display_name":"Skeleton-of-thought: Large language models can do parallel decoding","publication_year":2023,"publication_date":"2023-12-16","ids":{"openalex":"https://openalex.org/W7034070351"},"language":"en","primary_location":{"id":"pmh:oai:lirias2repo.kuleuven.be:20.500.12942/735245","doi":null,"is_oa":true,"landing_page_url":"https://lirias.kuleuven.be/handle/20.500.12942/735245","pdf_url":"https://lirias.kuleuven.be/retrieve/a5d96324-8966-4109-8baa-894d335916de","source":{"id":"https://openalex.org/S7407055369","display_name":"Lirias","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":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"3rd NeurIPS workshop on Efficient Natural Language and Speech Processing - ENLSP-III: Towards the Future of Large Language Models and their Emerging Descendants, New Orleans, USA, 16 December 2023","raw_type":"info:eu-repo/semantics/publishedVersion"},"volume":null,"type":"conference-paper","is_traditional":true,"indexed_in":[],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://lirias.kuleuven.be/retrieve/a5d96324-8966-4109-8baa-894d335916de","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Ning, Xuefei","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ning, Xuefei","raw_affiliation_strings":["Department of Electronic Engineering , Tsinghua University , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Electronic Engineering , Tsinghua University , Beijing , China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Lin, Zinan","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Lin, Zinan","raw_affiliation_strings":["Microsoft Research , Redmond , Washington , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft Research , Redmond , Washington , USA","institution_ids":["https://openalex.org/I1290206253"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Zhou, Zixuan","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhou, Zixuan","raw_affiliation_strings":["Department of Electronic Engineering , Tsinghua University , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Electronic Engineering , Tsinghua University , Beijing , China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5052362538","display_name":"Zifu Wang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I99464096","display_name":"KU Leuven","ror":"https://ror.org/05f950310","country_code":"BE","type":"education","lineage":["https://openalex.org/I99464096"]}],"countries":["BE"],"is_corresponding":false,"raw_author_name":"Wang, Zifu","raw_affiliation_strings":["ESAT-PSI , KU Leuven , Leuven , Belgium"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ESAT-PSI , KU Leuven , Leuven , Belgium","institution_ids":["https://openalex.org/I99464096"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Yang, Huazhong","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yang, Huazhong","raw_affiliation_strings":["Department of Electronic Engineering , Tsinghua University , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Electronic Engineering , Tsinghua University , Beijing , China","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"last","author":{"id":null,"display_name":"Wang, Yu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wang, Yu","raw_affiliation_strings":["Department of Electronic Engineering , Tsinghua University , Beijing , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Electronic Engineering , Tsinghua University , Beijing , China","institution_ids":["https://openalex.org/I99065089"]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.3314,"has_fulltext":true,"cited_by_count":3,"citation_normalized_percentile":{"value":0.53253921,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":89,"max":95},"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":true,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.7271000146865845,"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":0.7271000146865845,"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":0.2565999925136566,"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/T10286","display_name":"Information Retrieval and Search Behavior","score":0.002899999963119626,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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":0.890999972820282},{"id":"https://openalex.org/keywords/parallel-decoding","display_name":"parallel decoding","score":0.7889999747276306},{"id":"https://openalex.org/keywords/sequential-decoding","display_name":"sequential decoding","score":0.8610000014305115},{"id":"https://openalex.org/keywords/api-calls","display_name":"API calls","score":0.6549999713897705},{"id":"https://openalex.org/keywords/inference-efficiency","display_name":"inference efficiency","score":0.8259999752044678},{"id":"https://openalex.org/keywords/answer-quality","display_name":"answer quality","score":0.777999997138977},{"id":"https://openalex.org/keywords/gpt-4","display_name":"GPT-4","score":0.6679999828338623}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6869999766349792},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.515999972820282},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.40709999203681946},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.40470001101493835},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.4000999927520752},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3788999915122986},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.3368000090122223},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.3366999924182892},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.27630001306533813},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.27149999141693115},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.27079999446868896},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.2694000005722046},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2558000087738037},{"id":"https://openalex.org/C116834253","wikidata":"https://www.wikidata.org/wiki/Q2039217","display_name":"Identification (biology)","level":2,"score":0.2549999952316284},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.25360000133514404}],"mesh":[],"locations_count":1,"locations":[{"id":"pmh:oai:lirias2repo.kuleuven.be:20.500.12942/735245","doi":null,"is_oa":true,"landing_page_url":"https://lirias.kuleuven.be/handle/20.500.12942/735245","pdf_url":"https://lirias.kuleuven.be/retrieve/a5d96324-8966-4109-8baa-894d335916de","source":{"id":"https://openalex.org/S7407055369","display_name":"Lirias","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":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"3rd NeurIPS workshop on Efficient Natural Language and Speech Processing - ENLSP-III: Towards the Future of Large Language Models and their Emerging Descendants, New Orleans, USA, 16 December 2023","raw_type":"info:eu-repo/semantics/publishedVersion"}],"best_oa_location":{"id":"pmh:oai:lirias2repo.kuleuven.be:20.500.12942/735245","doi":null,"is_oa":true,"landing_page_url":"https://lirias.kuleuven.be/handle/20.500.12942/735245","pdf_url":"https://lirias.kuleuven.be/retrieve/a5d96324-8966-4109-8baa-894d335916de","source":{"id":"https://openalex.org/S7407055369","display_name":"Lirias","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":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"3rd NeurIPS workshop on Efficient Natural Language and Speech Processing - ENLSP-III: Towards the Future of Large Language Models and their Emerging Descendants, New Orleans, USA, 16 December 2023","raw_type":"info:eu-repo/semantics/publishedVersion"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320308943","display_name":"Microsoft Research","ror":"https://ror.org/00d0nc645"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7034070351.pdf","grobid_xml":"https://content.openalex.org/works/W7034070351.grobid-xml"},"referenced_works_count":38,"referenced_works":["https://openalex.org/W1598866093","https://openalex.org/W2119144962","https://openalex.org/W2809624076","https://openalex.org/W2969797940","https://openalex.org/W2988394319","https://openalex.org/W2994673210","https://openalex.org/W3033529678","https://openalex.org/W3152956381","https://openalex.org/W3156528192","https://openalex.org/W3163465001","https://openalex.org/W3174770825","https://openalex.org/W3198002980","https://openalex.org/W4221161695","https://openalex.org/W4304195432","https://openalex.org/W4307079201","https://openalex.org/W4307934016","https://openalex.org/W4309591680","https://openalex.org/W4309953112","https://openalex.org/W4310561894","https://openalex.org/W4319166707","https://openalex.org/W4322718191","https://openalex.org/W4324297016","https://openalex.org/W4327810158","https://openalex.org/W4367000491","https://openalex.org/W4378768661","https://openalex.org/W4379260375","https://openalex.org/W4384705353","https://openalex.org/W4385774107","https://openalex.org/W4387321091","https://openalex.org/W4389519291","https://openalex.org/W4389520103","https://openalex.org/W4389523706","https://openalex.org/W4393160302","https://openalex.org/W4395112660","https://openalex.org/W4406085118","https://openalex.org/W7133202489","https://openalex.org/W7133209886","https://openalex.org/W7133214447"],"related_works":["https://openalex.org/W7076001607","https://openalex.org/W7108657319","https://openalex.org/W4414981911","https://openalex.org/W4401043440","https://openalex.org/W7168283953","https://openalex.org/W7196944556","https://openalex.org/W7196931357","https://openalex.org/W7220806852","https://openalex.org/W7196941722","https://openalex.org/W4407713252"],"abstract_inverted_index":{"This":[0],"work":[1],"aims":[2],"at":[3,105],"decreasing":[4],"the":[5,15,19,24,38,55,58,71,94,114,142,154,165,172,194],"end-to-end":[6],"generation":[7,21],"latency":[8,22],"of":[9,14,18,43,57,73,116,147],"large":[10],"language":[11],"models":[12],"(LLMs).One":[13],"major":[16],"causes":[17],"high":[20],"is":[23,101],"sequential":[25],"decoding":[26,68],"approach":[27],"adopted":[28],"by":[29,37],"almost":[30],"all":[31],"state-of-the-art":[32],"LLMs.In":[33],"this":[34],"work,":[35],"motivated":[36],"thinking":[39],"and":[40,60,111,144],"writing":[41],"process":[42],"humans,":[44],"we":[45],"propose":[46],"Skeleton-of-Thought":[47],"(SoT),":[48],"which":[49,181],"first":[50],"guides":[51],"LLMs":[52,118],"to":[53,69,119,169,193],"generate":[54],"skeleton":[56,75],"answer,":[59],"then":[61],"conducts":[62],"parallel":[63],"API":[64],"calls":[65],"or":[66],"batched":[67],"complete":[70],"contents":[72],"each":[74],"point":[76],"in":[77,134],"parallel.Not":[78],"only":[79],"does":[80],"SoT":[81,148,173,184],"provide":[82],"considerable":[83],"speed-ups":[84],"across":[85],"12":[86],"LLMs,":[87],"but":[88],"it":[89],"can":[90],"also":[91],"potentially":[92],"improve":[93],"answer":[95,126,157,191],"quality":[96,143,158],"on":[97,149,185],"several":[98],"question":[99],"categories.SoT":[100],"an":[102],"initial":[103],"attempt":[104],"datacentric":[106],"optimization":[107],"for":[108,125,156],"inference":[109],"efficiency,":[110],"further":[112],"underscores":[113],"potential":[115],"pushing":[117],"think":[120],"more":[121],"like":[122],"a":[123],"human":[124],"quality.":[127],"*":[128],"Equal":[129],"contribution.\u2020":[130],"The":[131],"main":[132],"updates":[133],"arXiv":[135],"V2":[136],"are":[137,167],"as":[138,153,164],"follows:":[139],"(1)":[140],"Add":[141,171],"efficiency":[145],"evaluation":[146],"GPT-4.(2)":[150],"Use":[151],"GPT-4":[152],"judge":[155,166],"evaluation.The":[159],"old":[160],"results":[161],"with":[162,174],"ChatGPT-3.5":[163],"moved":[168],"App.I.3.(3)":[170],"Router":[175],"(SoT-R)":[176],"method":[177],"(":[178],"\u00a7":[179],"4)":[180,188],"adaptively":[182],"triggers":[183],"suitable":[186],"questions.(":[187],"Move":[189],"detailed":[190],"analysis":[192],"appendices.":[195]},"counts_by_year":[{"year":2025,"cited_by_count":1},{"year":2024,"cited_by_count":2}],"updated_date":"2026-11-08T12:12:28.789561","created_date":"2025-10-10T00:00:00"}