{"id":"https://openalex.org/W4412886673","doi":"https://doi.org/10.18653/v1/2025.acl-long.169","title":"Which of These Best Describes Multiple Choice Evaluation with LLMs? A) Forced B) Flawed C) Fixable D) All of the Above","display_name":"Which of These Best Describes Multiple Choice Evaluation with LLMs? A) Forced B) Flawed C) Fixable D) All of the Above","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412886673","doi":"https://doi.org/10.18653/v1/2025.acl-long.169"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.acl-long.169","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.169","pdf_url":"https://aclanthology.org/2025.acl-long.169.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":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.acl-long.169.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5091963084","display_name":"Nishant Balepur","orcid":null},"institutions":[{"id":"https://openalex.org/I66946132","display_name":"University of Maryland, College Park","ror":"https://ror.org/047s2c258","country_code":"US","type":"education","lineage":["https://openalex.org/I66946132"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Nishant Balepur","raw_affiliation_strings":["University of Maryland"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Maryland","institution_ids":["https://openalex.org/I66946132"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5082447472","display_name":"Rachel Rudinger","orcid":"https://orcid.org/0000-0002-5506-4701"},"institutions":[{"id":"https://openalex.org/I66946132","display_name":"University of Maryland, College Park","ror":"https://ror.org/047s2c258","country_code":"US","type":"education","lineage":["https://openalex.org/I66946132"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Rachel Rudinger","raw_affiliation_strings":["University of Maryland"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Maryland","institution_ids":["https://openalex.org/I66946132"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5092513269","display_name":"Jordan Lee Boyd-Graber","orcid":null},"institutions":[{"id":"https://openalex.org/I66946132","display_name":"University of Maryland, College Park","ror":"https://ror.org/047s2c258","country_code":"US","type":"education","lineage":["https://openalex.org/I66946132"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jordan Lee Boyd-Graber","raw_affiliation_strings":["University of Maryland"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Maryland","institution_ids":["https://openalex.org/I66946132"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I66946132"],"apc_list":null,"apc_paid":null,"fwci":32.9475,"has_fulltext":true,"cited_by_count":5,"citation_normalized_percentile":{"value":0.99531396,"is_in_top_1_percent":true,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"3394","last_page":"3418"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12755","display_name":"Legal Education and Practice Innovations","score":0.8198000192642212,"subfield":{"id":"https://openalex.org/subfields/3308","display_name":"Law"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},"topics":[{"id":"https://openalex.org/T12755","display_name":"Legal Education and Practice Innovations","score":0.8198000192642212,"subfield":{"id":"https://openalex.org/subfields/3308","display_name":"Law"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T13643","display_name":"Artificial Intelligence in Law","score":0.7347999811172485,"subfield":{"id":"https://openalex.org/subfields/3320","display_name":"Political Science and International Relations"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T14230","display_name":"Occupational and Professional Licensing Regulation","score":0.7081999778747559,"subfield":{"id":"https://openalex.org/subfields/2002","display_name":"Economics and Econometrics"},"field":{"id":"https://openalex.org/fields/20","display_name":"Economics, Econometrics and Finance"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/operations-research","display_name":"Operations research","score":0.3661677837371826},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.34971535205841064},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.219371497631073}],"concepts":[{"id":"https://openalex.org/C42475967","wikidata":"https://www.wikidata.org/wiki/Q194292","display_name":"Operations research","level":1,"score":0.3661677837371826},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.34971535205841064},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.219371497631073}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.acl-long.169","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.169","pdf_url":"https://aclanthology.org/2025.acl-long.169.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":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.acl-long.169","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.169","pdf_url":"https://aclanthology.org/2025.acl-long.169.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":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412886673.pdf","grobid_xml":"https://content.openalex.org/works/W4412886673.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"Multiple":[0],"choice":[1],"question":[2],"answering":[3],"(MCQA)":[4],"is":[5,75],"popular":[6],"for":[7,20,48,214],"LLM":[8,38,116,173],"evaluation":[9],"due":[10],"to":[11,68,97,103,110,138,164,280],"its":[12,21,79],"simplicity":[13],"and":[14,41,57,63,86,106,121],"human-like":[15],"testing,":[16,152],"but":[17],"we":[18,90,114,134,141],"argue":[19],"reform.We":[22],"first":[23],"reveal":[24],"flaws":[25],"in":[26,118,145],"MCQA's":[27,159],"format,":[28,78],"as":[29],"it":[30],"struggles":[31],"to:":[32],"1)":[33],"test":[34,44,182,259],"generation/subjectivity;":[35],"2)":[36],"match":[37],"use":[39,174],"cases;":[40],"3)":[42],"fully":[43,181],"knowledge.We":[45],"instead":[46],"advocate":[47],"generative":[49,253],"formats":[50,254],"based":[51,149],"on":[52,150],"human":[53],"testing-where":[54],"LLMs":[55,215,220],"construct":[56],"explain":[58],"answers-better":[59],"capturing":[60],"user":[61],"needs":[62],"knowledge":[64,183,250],"while":[65],"remaining":[66],"easy":[67,213],"score.We":[69],"then":[70],"show":[71],"even":[72],"when":[73],"MCQA":[74,189],"a":[76],"useful":[77],"datasets":[80],"suffer":[81],"from:":[82],"leakage;":[83],"unanswerability;":[84],"shortcuts;":[85],"saturation.In":[87],"each":[88],"issue,":[89],"give":[91,237],"fixes":[92],"from":[93,245],"education,":[94],"like":[95],"rubrics":[96,273],"guide":[98],"MCQ":[99],"writing;":[100],"scoring":[101,279],"methods":[102],"bridle":[104],"guessing;":[105],"Item":[107],"Response":[108],"Theory":[109],"build":[111],"harder":[112,286],"MCQs.Lastly,":[113],"discuss":[115],"errors":[117],"MCQA-robustness,":[119],"biases,":[120],"unfaithful":[122,238],"explanations-showing":[123],"how":[124],"our":[125],"prior":[126],"solutions":[127],"better":[128],"measure":[129],"or":[130],"address":[131],"these":[132],"issues.While":[133],"do":[135,219],"not":[136],"need":[137],"desert":[139],"MCQA,":[140],"encourage":[142],"more":[143],"efforts":[144],"refining":[146],"the":[147],"task":[148],"educational":[151,272],"advancing":[153],"evaluations.":[154],"Q1.":[155],"What's":[156],"wrong":[157,187],"with":[158,172,188,222,262,288],"format?A)":[160],"It":[161,179],"doesn't":[162,180],"apply":[163],"many":[165],"tasks":[166],"(":[167,176,184,195,202,208,216,227,233,240,255,265,274,283,292],"\u00a73.1)":[168],"B)":[169,197,229,257],"It's":[170],"misaligned":[171],"cases":[175],"\u00a73.2)":[177],"C)":[178,204,235,267],"\u00a73.3)":[185],"Q2.What's":[186],"datasets?A)":[190],"Test":[191],"sets":[192],"are":[193],"contaminated":[194],"\u00a75.1)":[196,266],"They":[198,205,224,230,236],"have":[199],"unanswerable":[200],"questions":[201,264],"\u00a75.2)":[203,275],"contain":[206],"shortcuts":[207],"\u00a75.3)":[209],"D)":[210,276],"They're":[211],"too":[212],"\u00a75.4)":[217],"Q3.How":[218],"struggle":[221],"MCQA?A)":[223,248],"lack":[225],"robustness":[226],"\u00a76.1)":[228],"exhibit":[231],"biases":[232],"\u00a76.2)":[234],"explanations":[239],"\u00a76.3)":[241],"Q4.How":[242],"can":[243],"insights":[244],"education":[246],"improve":[247],"Improve":[249],"testing":[251],"via":[252],"\u00a74)":[256],"Combat":[258],"set":[260],"leakage":[261],"fresh":[263],"Write":[268],"MCQs":[269,287],"informed":[270],"by":[271],"Use":[277],"calibration":[278],"curb":[281],"guessing":[282],"\u00a75.3.1)E)":[284],"Find":[285],"item":[289],"response":[290],"theory":[291],"\u00a75.4.1)":[293]},"counts_by_year":[{"year":2026,"cited_by_count":5}],"updated_date":"2026-08-21T09:56:20.448147","created_date":"2025-10-10T00:00:00"}