{"id":"https://openalex.org/W4403344710","doi":"https://doi.org/10.48550/arxiv.2410.06238","title":"EVOLvE: Evaluating and Optimizing LLMs For In-Context Exploration","display_name":"EVOLvE: Evaluating and Optimizing LLMs For In-Context Exploration","publication_year":2024,"publication_date":"2024-10-08","ids":{"openalex":"https://openalex.org/W4403344710","doi":"https://doi.org/10.48550/arxiv.2410.06238"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2410.06238","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2410.06238","pdf_url":"https://arxiv.org/pdf/2410.06238","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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/2410.06238","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5018535992","display_name":"Allen Nie","orcid":"https://orcid.org/0000-0001-6483-2111","observed_orcids":["https://orcid.org/0000-0001-6483-2111"]},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Nie, Allen","raw_affiliation_strings":["Stanford University , 2 Google DeepMind ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Stanford University , 2 Google DeepMind ,","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5004835193","display_name":"Yi Su","orcid":"https://orcid.org/0000-0002-1946-8063","observed_orcids":["https://orcid.org/0000-0002-1946-8063"]},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Su, Yi","raw_affiliation_strings":["Stanford University , 2 Google DeepMind ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Stanford University , 2 Google DeepMind ,","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5030537843","display_name":"Bo Chang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Chang, Bo","raw_affiliation_strings":["Stanford University , 2 Google DeepMind ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Stanford University , 2 Google DeepMind ,","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100348028","display_name":"Jonathan Lee","orcid":"https://orcid.org/0000-0002-1828-1707","observed_orcids":["https://orcid.org/0000-0002-1828-1707"]},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Lee, Jonathan N.","raw_affiliation_strings":["Stanford University , 2 Google DeepMind ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Stanford University , 2 Google DeepMind ,","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5028125399","display_name":"Ed H.","orcid":"https://orcid.org/0000-0003-3230-5338","observed_orcids":["https://orcid.org/0000-0003-3230-5338"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chi, Ed H.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5088551093","display_name":"Quoc Viet Le","orcid":"https://orcid.org/0000-0002-1087-2844","observed_orcids":["https://orcid.org/0000-0002-1087-2844"]},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Le, Quoc V.","raw_affiliation_strings":["Stanford University , 2 Google DeepMind ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Stanford University , 2 Google DeepMind ,","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100699702","display_name":"Minmin Chen","orcid":"https://orcid.org/0000-0002-7342-9022","observed_orcids":["https://orcid.org/0000-0002-7342-9022"]},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Chen, Minmin","raw_affiliation_strings":["Stanford University , 2 Google DeepMind ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Stanford University , 2 Google DeepMind ,","institution_ids":["https://openalex.org/I97018004"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":1,"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/T12101","display_name":"Advanced Bandit Algorithms Research","score":0.9994999766349792,"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"}},"topics":[{"id":"https://openalex.org/T12101","display_name":"Advanced Bandit Algorithms Research","score":0.9994999766349792,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.0003000000142492354,"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/T12072","display_name":"Machine Learning and Algorithms","score":0.0,"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/large-language-models","display_name":"large language models","score":0.9779999852180481},{"id":"https://openalex.org/keywords/exploration","display_name":"exploration","score":0.3959999978542328},{"id":"https://openalex.org/keywords/contextual-bandits","display_name":"contextual bandits","score":0.828000009059906},{"id":"https://openalex.org/keywords/bandit-learning","display_name":"bandit learning","score":0.4959999918937683},{"id":"https://openalex.org/keywords/in-context-learning","display_name":"in-context learning","score":0.8659999966621399},{"id":"https://openalex.org/keywords/regret-analysis","display_name":"regret analysis","score":0.5879999995231628},{"id":"https://openalex.org/keywords/decision-making-under-uncertainty","display_name":"decision making under uncertainty","score":0.796999990940094},{"id":"https://openalex.org/keywords/synthetic-data","display_name":"synthetic data","score":0.5170000195503235},{"id":"https://openalex.org/keywords/model-size","display_name":"model size","score":0.47200000286102295},{"id":"https://openalex.org/keywords/llm-benchmarking","display_name":"LLM benchmarking","score":0.6700000166893005}],"concepts":[{"id":"https://openalex.org/C175605778","wikidata":"https://www.wikidata.org/wiki/Q3299701","display_name":"Natural resource economics","level":1,"score":0.3670099377632141},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"score":0.3627392649650574},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.3230167031288147},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.2748819589614868}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2410.06238","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2410.06238","pdf_url":"https://arxiv.org/pdf/2410.06238","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2410.06238","doi":"https://doi.org/10.48550/arxiv.2410.06238","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2410.06238","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2410.06238","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2410.06238","pdf_url":"https://arxiv.org/pdf/2410.06238","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4403344710.pdf","grobid_xml":"https://content.openalex.org/works/W4403344710.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4407123300","https://openalex.org/W4415795297","https://openalex.org/W7083801040","https://openalex.org/W4417172997","https://openalex.org/W4391505508","https://openalex.org/W4416365907","https://openalex.org/W7151881015","https://openalex.org/W4414634790","https://openalex.org/W7203692970","https://openalex.org/W7110463605"],"abstract_inverted_index":{"Despite":[0],"their":[1],"success":[2],"in":[3,12,60],"many":[4,23,69],"domains,":[5],"large":[6],"language":[7],"models":[8,149],"(LLMs)":[9],"remain":[10],"under-studied":[11],"scenarios":[13],"requiring":[14],"optimal":[15,45,58,97],"decision-making":[16],"under":[17],"uncertainty.":[18],"This":[19],"is":[20],"crucial":[21],"as":[22,166],"real-world":[24],"applications,":[25],"ranging":[26],"from":[27,131],"personalized":[28],"recommendations":[29],"to":[30,43,56,68,88,104,139,159,198,200],"healthcare":[31],"interventions,":[32],"demand":[33],"that":[34,172],"LLMs":[35],"not":[36],"only":[37],"predict":[38],"but":[39],"also":[40],"actively":[41],"learn":[42],"make":[44,57],"decisions":[46,59],"through":[47,119],"exploration.":[48,178],"In":[49],"this":[50,106],"work,":[51],"we":[52,100,180],"measure":[53],"LLMs'":[54,90],"(in)ability":[55],"bandits,":[61],"a":[62,73,182],"state-less":[63],"reinforcement":[64],"learning":[65],"setting":[66],"relevant":[67],"applications.":[70],"We":[71,153],"develop":[72],"comprehensive":[74],"suite":[75],"of":[76,96,176,185,193],"environments,":[77],"including":[78],"both":[79],"context-free":[80],"and":[81,118,125,169,204],"contextual":[82],"bandits":[83],"with":[84,144],"varying":[85],"task":[86,167],"difficulties,":[87],"benchmark":[89],"performance.":[91],"Motivated":[92],"by":[93,111],"the":[94,174,186,191,201],"existence":[95],"exploration":[98,142,188],"algorithms,":[99],"propose":[101],"efficient":[102],"ways":[103],"integrate":[105],"algorithmic":[107],"knowledge":[108],"into":[109],"LLMs:":[110],"providing":[112],"explicit":[113],"algorithm-guided":[114],"support":[115],"during":[116],"inference;":[117],"algorithm":[120],"distillation":[121],"via":[122],"in-context":[123],"demonstrations":[124],"fine-tuning,":[126],"using":[127,190],"synthetic":[128],"data":[129,170],"generated":[130],"these":[132,135],"algorithms.":[133],"Impressively,":[134],"techniques":[136],"allow":[137],"us":[138],"achieve":[140],"superior":[141],"performance":[143],"smaller":[145],"models,":[146],"surpassing":[147],"larger":[148],"on":[150,162],"various":[151,163],"tasks.":[152],"conducted":[154],"an":[155],"extensive":[156],"ablation":[157],"study":[158],"shed":[160],"light":[161],"factors,":[164],"such":[165],"difficulty":[168],"representation,":[171],"influence":[173],"efficiency":[175,189],"LLM":[177],"Additionally,":[179],"conduct":[181],"rigorous":[183],"analysis":[184],"LLM's":[187],"concept":[192],"regret,":[194],"linking":[195],"its":[196],"ability":[197],"explore":[199],"model":[202],"size":[203],"underlying":[205],"algorithm.":[206]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-09-30T09:27:27.532063","created_date":"2024-10-12T00:00:00"}