{"id":"https://openalex.org/W7166798349","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1031","title":"EvoCoT: Overcoming the Exploration Bottleneck in Reinforcement Learning for LLMs","display_name":"EvoCoT: Overcoming the Exploration Bottleneck in Reinforcement Learning for LLMs","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166798349","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1031"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.findings-acl.1031","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1031","pdf_url":"https://aclanthology.org/2026.findings-acl.1031.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: ACL 2026","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.findings-acl.1031.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139772658","display_name":"Huanyu Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huanyu Liu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139792842","display_name":"Jia Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jia Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139807747","display_name":"Yihong Dong","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yihong Dong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139772030","display_name":"Chang Yu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chang Yu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139770907","display_name":"Taozhi Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Taozhi Chen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139839899","display_name":"Lecheng Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lecheng Wang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134144978","display_name":"Yongding Tao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yongding Tao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139830773","display_name":"Bin Gu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bin Gu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139845338","display_name":"Ge Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ge Li","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":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.78750399,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"20589","last_page":"20603"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10462","display_name":"Reinforcement Learning in Robotics","score":0.7756999731063843,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.7756999731063843,"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/T10653","display_name":"Robot Manipulation and Learning","score":0.02969999983906746,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11975","display_name":"Evolutionary Algorithms and Applications","score":0.022299999371170998,"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/reinforcement-learning","display_name":"Reinforcement learning","score":0.5612999796867371},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.5141000151634216},{"id":"https://openalex.org/keywords/control","display_name":"Control (management)","score":0.37220001220703125},{"id":"https://openalex.org/keywords/reinforcement","display_name":"Reinforcement","score":0.3353999853134155},{"id":"https://openalex.org/keywords/action","display_name":"Action (physics)","score":0.28790000081062317},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.2777000069618225}],"concepts":[{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5612999796867371},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.5141000151634216},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4925999939441681},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.37220001220703125},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.35850000381469727},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.3463999927043915},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3370000123977661},{"id":"https://openalex.org/C67203356","wikidata":"https://www.wikidata.org/wiki/Q1321905","display_name":"Reinforcement","level":2,"score":0.3353999853134155},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.28790000081062317},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2777000069618225},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.27320000529289246},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.26980000734329224},{"id":"https://openalex.org/C136197465","wikidata":"https://www.wikidata.org/wiki/Q1729295","display_name":"Variety (cybernetics)","level":2,"score":0.26910001039505005},{"id":"https://openalex.org/C2778137410","wikidata":"https://www.wikidata.org/wiki/Q2732820","display_name":"Government (linguistics)","level":2,"score":0.260699987411499},{"id":"https://openalex.org/C539667460","wikidata":"https://www.wikidata.org/wiki/Q2414942","display_name":"Management science","level":1,"score":0.2572000026702881},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.25380000472068787},{"id":"https://openalex.org/C42475967","wikidata":"https://www.wikidata.org/wiki/Q194292","display_name":"Operations research","level":1,"score":0.25290000438690186}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.findings-acl.1031","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1031","pdf_url":"https://aclanthology.org/2026.findings-acl.1031.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: ACL 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-acl.1031","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1031","pdf_url":"https://aclanthology.org/2026.findings-acl.1031.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: ACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G3446613798","display_name":null,"funder_award_id":"62192730","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G5482656022","display_name":null,"funder_award_id":"62192731","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G8449925957","display_name":null,"funder_award_id":"62192733","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/W7166798349.pdf","grobid_xml":"https://content.openalex.org/works/W7166798349.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Reinforcement":[0],"learning":[1,36,68],"with":[2,145],"verifiable":[3],"reward":[4,32],"(RLVR)":[5],"has":[6],"become":[7],"a":[8,65,97],"promising":[9],"paradigm":[10],"for":[11,48],"post-training":[12],"large":[13],"language":[14],"models":[15,47,155],"(LLMs)":[16],"to":[17,92,103,116,130,156],"improve":[18],"their":[19],"reasoning":[20,59,75,136],"capability.However,":[21],"when":[22],"the":[23,31,78,94,151],"rollout":[24],"accuracy":[25],"is":[26,143],"low":[27],"on":[28,45,71],"hard":[29,109],"problems,":[30,54,134],"becomes":[33],"sparse,":[34],"limiting":[35,55],"efficiency":[37],"and":[38,83,123,142,154],"causing":[39],"exploration":[40,79],"bottlenecks.Existing":[41],"approaches":[42],"either":[43],"rely":[44],"teacher":[46],"distillation":[49],"or":[50,57],"filter":[51],"out":[52],"difficult":[53],"scalability":[56],"restricting":[58],"improvement":[60],"through":[61],"exploration.We":[62],"propose":[63],"EvoCoT,":[64],"self-Evolving":[66],"curriculum":[67],"framework":[69,100],"based":[70],"twostage":[72],"Chain-of-Thought":[73],"(CoT)":[74],"optimization.EvoCoT":[76],"constrains":[77],"space":[80,95],"by":[81],"self-generating":[82],"verifying":[84],"CoT":[85,90,140],"trajectories,":[86],"then":[87],"gradually":[88],"shortens":[89],"steps":[91],"expand":[93],"in":[96],"controlled":[98],"way.The":[99],"enables":[101,128],"LLMs":[102,129],"stably":[104],"learn":[105],"from":[106],"initially":[107],"unsolved":[108,133],"problems":[110],"under":[111],"sparse":[112],"rewards.We":[113],"apply":[114],"EvoCoT":[115,127],"multiple":[117],"LLM":[118],"families,":[119],"including":[120],"Qwen,":[121],"DeepSeek,":[122],"Llama.Experiments":[124],"show":[125],"that":[126],"solve":[131],"previously":[132],"improves":[135],"capability":[137],"without":[138],"external":[139],"supervision,":[141],"compatible":[144],"various":[146],"RL":[147],"fine-tuning":[148],"methods.We":[149],"release":[150],"source":[152],"code":[153],"support":[157],"future":[158],"at":[159],"https://github.com/gtxygyzb/EvoCoT.":[160]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2026-07-02T00:00:00"}