{"id":"https://openalex.org/W7127316352","doi":"https://doi.org/10.48550/arxiv.2602.01711","title":"Optimizing Prompts for Large Language Models: A Causal Approach","display_name":"Optimizing Prompts for Large Language Models: A Causal Approach","publication_year":2026,"publication_date":"2026-02-02","ids":{"openalex":"https://openalex.org/W7127316352","doi":"https://doi.org/10.48550/arxiv.2602.01711"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2602.01711","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.01711","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2602.01711","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5124882368","display_name":"Wei Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Chen, Wei","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124936751","display_name":"Yanbin Fang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fang, Yanbin","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101266068","display_name":"Shuran Fu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fu, Shuran","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124931045","display_name":"Fasheng Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Fasheng","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5124935028","display_name":"Xuan Wei","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wei, Xuan","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5124882368"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"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/T11948","display_name":"Machine Learning in Materials Science","score":0.18209999799728394,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11948","display_name":"Machine Learning in Materials Science","score":0.18209999799728394,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12535","display_name":"Machine Learning and Data Classification","score":0.06790000200271606,"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/T11273","display_name":"Advanced Graph Neural Networks","score":0.0608999989926815,"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/causal-inference","display_name":"Causal inference","score":0.7760000228881836},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.5924999713897705},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5866000056266785},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.5625},{"id":"https://openalex.org/keywords/personalization","display_name":"Personalization","score":0.508400022983551},{"id":"https://openalex.org/keywords/causal-model","display_name":"Causal model","score":0.4832000136375427},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.3001999855041504}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.815500020980835},{"id":"https://openalex.org/C158600405","wikidata":"https://www.wikidata.org/wiki/Q5054566","display_name":"Causal inference","level":2,"score":0.7760000228881836},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.5924999713897705},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5866000056266785},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5737000107765198},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.5625},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5533999800682068},{"id":"https://openalex.org/C183003079","wikidata":"https://www.wikidata.org/wiki/Q1000371","display_name":"Personalization","level":2,"score":0.508400022983551},{"id":"https://openalex.org/C11671645","wikidata":"https://www.wikidata.org/wiki/Q5054567","display_name":"Causal model","level":2,"score":0.4832000136375427},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3001999855041504},{"id":"https://openalex.org/C2780102126","wikidata":"https://www.wikidata.org/wiki/Q10928179","display_name":"Online and offline","level":2,"score":0.2736000120639801},{"id":"https://openalex.org/C2777402240","wikidata":"https://www.wikidata.org/wiki/Q6783436","display_name":"Masking (illustration)","level":2,"score":0.2718000113964081},{"id":"https://openalex.org/C149629883","wikidata":"https://www.wikidata.org/wiki/Q660926","display_name":"Fraction (chemistry)","level":2,"score":0.26829999685287476},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.26030001044273376},{"id":"https://openalex.org/C77350462","wikidata":"https://www.wikidata.org/wiki/Q1125472","display_name":"Confounding","level":2,"score":0.2590999901294708},{"id":"https://openalex.org/C2986087404","wikidata":"https://www.wikidata.org/wiki/Q15946010","display_name":"Online learning","level":2,"score":0.2565999925136566},{"id":"https://openalex.org/C163504300","wikidata":"https://www.wikidata.org/wiki/Q2364925","display_name":"Causal structure","level":2,"score":0.2508000135421753},{"id":"https://openalex.org/C137836250","wikidata":"https://www.wikidata.org/wiki/Q984063","display_name":"Optimization problem","level":2,"score":0.25}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2602.01711","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.01711","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2602.01711","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.01711","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs)":[3],"are":[4,66,174],"increasingly":[5],"embedded":[6],"in":[7,95,156,243],"enterprise":[8,244],"workflows,":[9],"yet":[10],"their":[11],"performance":[12],"remains":[13],"highly":[14],"sensitive":[15],"to":[16,24,51,53,112,137,187,206],"prompt":[17,38,70,85,124,197,241],"design.":[18],"Automatic":[19],"Prompt":[20,78],"Optimization":[21,79],"(APO)":[22],"seeks":[23],"mitigate":[25],"this":[26,133],"instability,":[27],"but":[28,49],"existing":[29,184],"approaches":[30,59],"face":[31],"two":[32,96],"persistent":[33],"challenges.":[34],"First,":[35,98],"commonly":[36],"used":[37],"strategies":[39],"rely":[40],"on":[41,47,61,147,180],"static":[42],"instructions":[43],"that":[44,65,83],"perform":[45],"well":[46],"average":[48],"fail":[50],"adapt":[52],"heterogeneous":[54],"queries.":[55],"Second,":[56,130],"more":[57],"dynamic":[58],"depend":[60],"offline":[62,102,208],"reward":[63,104,135],"models":[64],"fundamentally":[67,192],"correlational,":[68],"confounding":[69,127],"effectiveness":[71],"with":[72],"query":[73,128],"characteristics.":[74],"We":[75,151],"propose":[76],"Causal":[77],"(CPO),":[80],"a":[81,88,139,217,234],"framework":[82],"reframes":[84],"design":[86],"as":[87,233],"problem":[89],"of":[90,115,123,196,219],"causal":[91,103,121,209,231],"estimation.":[92],"CPO":[93,153,163,191],"operates":[94],"stages.":[97],"it":[99,131,211],"learns":[100],"an":[101,207],"model":[105,204],"by":[106,177,199,224],"applying":[107],"Double":[108],"Machine":[109],"Learning":[110],"(DML)":[111],"semantic":[113],"embeddings":[114],"prompts":[116,144,167],"and":[117,160,168,239],"queries,":[118,182],"isolating":[119],"the":[120,194,220],"effect":[122],"variations":[125],"from":[126,202],"attributes.":[129],"utilizes":[132],"unbiased":[134],"signal":[136],"guide":[138],"resource-efficient":[140],"search":[141],"for":[142,237],"query-specific":[143],"without":[145],"relying":[146],"costly":[148],"online":[149,225],"evaluation.":[150],"evaluate":[152],"across":[154],"benchmarks":[155],"mathematical":[157],"reasoning,":[158],"visualization,":[159],"data":[161],"analytics.":[162],"consistently":[164],"outperforms":[165],"human-engineered":[166],"state-of-the-art":[169],"automated":[170],"optimizers.":[171],"The":[172],"gains":[173],"driven":[175],"primarily":[176],"improved":[178],"robustness":[179],"hard":[181],"where":[183],"methods":[185],"tend":[186],"deteriorate.":[188],"Beyond":[189],"performance,":[190],"reshapes":[193],"economics":[195],"optimization:":[198],"shifting":[200],"evaluation":[201],"real-time":[203],"execution":[205],"model,":[210],"enables":[212],"high-precision,":[213],"per-query":[214],"customization":[215],"at":[216],"fraction":[218],"inference":[221,232],"cost":[222],"required":[223],"methods.":[226],"Together,":[227],"these":[228],"results":[229],"establish":[230],"scalable":[235],"foundation":[236],"reliable":[238],"cost-efficient":[240],"optimization":[242],"LLM":[245],"deployments.":[246]},"counts_by_year":[],"updated_date":"2026-02-04T23:14:21.375766","created_date":"2026-02-04T00:00:00"}