{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/132610"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/132610","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Temperature-centric investigation of speculative decoding with knowledge distillation","abstract":"Speculative decoding stands as a pivotal technique to expedite inference in autoregressive (large) language models. This method employs a smaller \\textit{draft} model to speculate a block of tokens, which the \\textit{target} model then evaluates for acceptance. Despite a wealth of studies aimed at increasing the efficiency of speculative decoding, the influence of generation configurations on the decoding process remains poorly understood, especially concerning decoding temperatures. This paper delves into the effects of decoding temperatures on speculative decoding’s efficacy. Beginning with knowledge distillation (KD), we first highlight the challenge of decoding at higher temperatures, and demonstrate KD in a consistent temperature setting could be a remedy. We also investigate the effects of out-of-domain testing sets with out-of-range temperatures. Building upon these findings, we take an initial step to further the speedup for speculative decoding, particularly in a high-temperature generation setting. Our work offers new insights into how generation configurations drastically affect the performance of speculative decoding, and underscores the need for developing methods that focus on diverse decoding configurations. Code is publicly available at \\texttt{\\url{https://github.com/ozyyshr/TempSpec}}.","abstract_html":"Speculative decoding stands as a pivotal technique to expedite inference in autoregressive (large) language models. This method employs a smaller \\textit{draft} model to speculate a block of tokens, which the \\textit{target} model then evaluates for acceptance. Despite a wealth of studies aimed at increasing the efficiency of speculative decoding, the influence of generation configurations on the decoding process remains poorly understood, especially concerning decoding temperatures. This paper delves into the effects of decoding temperatures on speculative decoding’s efficacy. Beginning with knowledge distillation (KD), we first highlight the challenge of decoding at higher temperatures, and demonstrate KD in a consistent temperature setting could be a remedy. We also investigate the effects of out-of-domain testing sets with out-of-range temperatures. Building upon these findings, we take an initial step to further the speedup for speculative decoding, particularly in a high-temperature generation setting. Our work offers new insights into how generation configurations drastically affect the performance of speculative decoding, and underscores the need for developing methods that focus on diverse decoding configurations. Code is publicly available at \\texttt{\\url{https://github.com/ozyyshr/TempSpec}}.","abstract_has_math":false,"creators":["Ouyang, Siru"],"institution":"University of Illinois Urbana-Champaign","degree_name":"M.S.","degree_level":"Thesis","degree_discipline":"Computer Science","degree_department":null,"school":null,"contributors":["Han, Jiawei"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025-12","date_published":"2025-12","updated_at":"2026-07-22T22:25:07Z","subjects":["Speculative Decoding","Knowledge Distillation","Large Language Models"],"languages":["en"],"rights":["Copyright 2025 Siru Ouyang"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/2142/132610","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Han, Jiawei"]},{"key":"dc:creator","label":"Author","values":["Ouyang, Siru"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2025-12","2025-04-10"]},{"key":"dc:type","label":"Dc Type","values":["text","Thesis"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Computer Science"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Thesis"]},{"key":"thesis:degree_name","label":"Degree Name","values":["M.S."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Illinois Urbana-Champaign"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Speculative Decoding","Knowledge Distillation","Large Language Models"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2025 Siru Ouyang"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://hdl.handle.net/2142/132610"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Speculative decoding stands as a pivotal technique to expedite inference in autoregressive (large) language models. This method employs a smaller \\textit{draft} model to speculate a block of tokens, which the \\textit{target} model then evaluates for acceptance. Despite a wealth of studies aimed at increasing the efficiency of speculative decoding, the influence of generation configurations on the decoding process remains poorly understood, especially concerning decoding temperatures. This paper delves into the effects of decoding temperatures on speculative decoding’s efficacy. Beginning with knowledge distillation (KD), we first highlight the challenge of decoding at higher temperatures, and demonstrate KD in a consistent temperature setting could be a remedy. We also investigate the effects of out-of-domain testing sets with out-of-range temperatures. Building upon these findings, we take an initial step to further the speedup for speculative decoding, particularly in a high-temperature generation setting. Our work offers new insights into how generation configurations drastically affect the performance of speculative decoding, and underscores the need for developing methods that focus on diverse decoding configurations. Code is publicly available at \\texttt{\\url{https://github.com/ozyyshr/TempSpec}}.","Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2027-12-01","The student, Siru Ouyang, accepted the attached license on 2025-04-10 at 11:48.","The student, Siru Ouyang, submitted this Thesis for approval on 2025-04-10 at 11:53.","This Thesis was approved for publication on 2025-04-10 at 13:13.","DSpace SAF Submission Ingestion Package generated from Vireo submission #21738 on 2026-02-19 at 18:45:08"]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Temperature-centric investigation of speculative decoding with knowledge distillation"]}]}],"canonical_facts":{"dc:contributor":["Han, Jiawei"],"dc:creator":["Ouyang, Siru"],"dc:date":["2025-12","2025-04-10"],"dc:description":["Speculative decoding stands as a pivotal technique to expedite inference in autoregressive (large) language models. This method employs a smaller \\textit{draft} model to speculate a block of tokens, which the \\textit{target} model then evaluates for acceptance. Despite a wealth of studies aimed at increasing the efficiency of speculative decoding, the influence of generation configurations on the decoding process remains poorly understood, especially concerning decoding temperatures. This paper delves into the effects of decoding temperatures on speculative decoding’s efficacy. Beginning with knowledge distillation (KD), we first highlight the challenge of decoding at higher temperatures, and demonstrate KD in a consistent temperature setting could be a remedy. We also investigate the effects of out-of-domain testing sets with out-of-range temperatures. Building upon these findings, we take an initial step to further the speedup for speculative decoding, particularly in a high-temperature generation setting. Our work offers new insights into how generation configurations drastically affect the performance of speculative decoding, and underscores the need for developing methods that focus on diverse decoding configurations. Code is publicly available at \\texttt{\\url{https://github.com/ozyyshr/TempSpec}}.","Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2027-12-01","The student, Siru Ouyang, accepted the attached license on 2025-04-10 at 11:48.","The student, Siru Ouyang, submitted this Thesis for approval on 2025-04-10 at 11:53.","This Thesis was approved for publication on 2025-04-10 at 13:13.","DSpace SAF Submission Ingestion Package generated from Vireo submission #21738 on 2026-02-19 at 18:45:08"],"dc:format":["application/pdf"],"dc:identifier":["https://hdl.handle.net/2142/132610"],"dc:language":["en"],"dc:rights":["Copyright 2025 Siru Ouyang"],"dc:subject":["Speculative Decoding","Knowledge Distillation","Large Language Models"],"dc:title":["Temperature-centric investigation of speculative decoding with knowledge distillation"],"dc:type":["text","Thesis"],"thesis:degree_discipline":["Computer Science"],"thesis:degree_level":["Thesis"],"thesis:degree_name":["M.S."],"thesis:institution_name":["University of Illinois Urbana-Champaign"]},"updated_at":"2026-07-22T22:25:07Z"}