{"id":"https://openalex.org/W4389518636","doi":"https://doi.org/10.18653/v1/2023.findings-emnlp.973","title":"Emergent Inabilities? Inverse Scaling Over the Course of Pretraining","display_name":"Emergent Inabilities? Inverse Scaling Over the Course of Pretraining","publication_year":2023,"publication_date":"2023-01-01","ids":{"openalex":"https://openalex.org/W4389518636","doi":"https://doi.org/10.18653/v1/2023.findings-emnlp.973"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2023.findings-emnlp.973","is_oa":true,"landing_page_url":"http://dx.doi.org/10.18653/v1/2023.findings-emnlp.973","pdf_url":"https://aclanthology.org/2023.findings-emnlp.973.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: EMNLP 2023","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2023.findings-emnlp.973.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5054612798","display_name":"James A. Michaelov","orcid":"https://orcid.org/0000-0003-2913-1103"},"institutions":[{"id":"https://openalex.org/I36258959","display_name":"University of California San Diego","ror":"https://ror.org/0168r3w48","country_code":"US","type":"education","lineage":["https://openalex.org/I36258959"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"James Michaelov","raw_affiliation_strings":["Department of Cognitive Science, University of California San Diego"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Cognitive Science, University of California San Diego","institution_ids":["https://openalex.org/I36258959"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5043344696","display_name":"Benjamin K. Bergen","orcid":"https://orcid.org/0000-0002-9395-9151"},"institutions":[{"id":"https://openalex.org/I36258959","display_name":"University of California San Diego","ror":"https://ror.org/0168r3w48","country_code":"US","type":"education","lineage":["https://openalex.org/I36258959"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Ben Bergen","raw_affiliation_strings":["Department of Cognitive Science, University of California San Diego"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Cognitive Science, University of California San Diego","institution_ids":["https://openalex.org/I36258959"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":["https://openalex.org/A5043344696"],"corresponding_institution_ids":["https://openalex.org/I36258959"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.19134842,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"14607","last_page":"14615"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9998000264167786,"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/T10028","display_name":"Topic Modeling","score":0.9998000264167786,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.9973999857902527,"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/T13629","display_name":"Text Readability and Simplification","score":0.9822999835014343,"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/hindsight-bias","display_name":"Hindsight bias","score":0.8330224752426147},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.696453332901001},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.6895191669464111},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.6098291873931885},{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.6043370962142944},{"id":"https://openalex.org/keywords/inverse","display_name":"Inverse","score":0.444416880607605},{"id":"https://openalex.org/keywords/task-analysis","display_name":"Task analysis","score":0.441620796918869},{"id":"https://openalex.org/keywords/function","display_name":"Function (biology)","score":0.42372092604637146},{"id":"https://openalex.org/keywords/neglect","display_name":"Neglect","score":0.4204578399658203},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3908461332321167},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.382654070854187},{"id":"https://openalex.org/keywords/cognitive-psychology","display_name":"Cognitive psychology","score":0.38158857822418213},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.3452046513557434},{"id":"https://openalex.org/keywords/psychology","display_name":"Psychology","score":0.19670426845550537},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.11624425649642944},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.07841339707374573}],"concepts":[{"id":"https://openalex.org/C10347200","wikidata":"https://www.wikidata.org/wiki/Q1960297","display_name":"Hindsight bias","level":2,"score":0.8330224752426147},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.696453332901001},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.6895191669464111},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.6098291873931885},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.6043370962142944},{"id":"https://openalex.org/C207467116","wikidata":"https://www.wikidata.org/wiki/Q4385666","display_name":"Inverse","level":2,"score":0.444416880607605},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.441620796918869},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.42372092604637146},{"id":"https://openalex.org/C2776289891","wikidata":"https://www.wikidata.org/wiki/Q1931511","display_name":"Neglect","level":2,"score":0.4204578399658203},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3908461332321167},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.382654070854187},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.38158857822418213},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3452046513557434},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.19670426845550537},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.11624425649642944},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.07841339707374573},{"id":"https://openalex.org/C201995342","wikidata":"https://www.wikidata.org/wiki/Q682496","display_name":"Systems engineering","level":1,"score":0.0},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.0},{"id":"https://openalex.org/C78458016","wikidata":"https://www.wikidata.org/wiki/Q840400","display_name":"Evolutionary biology","level":1,"score":0.0},{"id":"https://openalex.org/C118552586","wikidata":"https://www.wikidata.org/wiki/Q7867","display_name":"Psychiatry","level":1,"score":0.0},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2023.findings-emnlp.973","is_oa":true,"landing_page_url":"http://dx.doi.org/10.18653/v1/2023.findings-emnlp.973","pdf_url":"https://aclanthology.org/2023.findings-emnlp.973.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: EMNLP 2023","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2023.findings-emnlp.973","is_oa":true,"landing_page_url":"http://dx.doi.org/10.18653/v1/2023.findings-emnlp.973","pdf_url":"https://aclanthology.org/2023.findings-emnlp.973.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: EMNLP 2023","raw_type":"proceedings-article"},"sustainable_development_goals":[{"score":0.44999998807907104,"id":"https://metadata.un.org/sdg/10","display_name":"Reduced inequalities"}],"awards":[],"funders":[{"id":"https://openalex.org/F4320309480","display_name":"Nvidia","ror":"https://ror.org/03jdj4y14"}],"has_content":{"grobid_xml":false,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4389518636.pdf"},"referenced_works_count":24,"referenced_works":["https://openalex.org/W2110065044","https://openalex.org/W3001279689","https://openalex.org/W3118781290","https://openalex.org/W3121904249","https://openalex.org/W3133702157","https://openalex.org/W3168194750","https://openalex.org/W3201174429","https://openalex.org/W4200634402","https://openalex.org/W4224308101","https://openalex.org/W4226146865","https://openalex.org/W4226153346","https://openalex.org/W4226399820","https://openalex.org/W4281690148","https://openalex.org/W4283026156","https://openalex.org/W4286850188","https://openalex.org/W4292779060","https://openalex.org/W4296413526","https://openalex.org/W4297412056","https://openalex.org/W4320005767","https://openalex.org/W4322718191","https://openalex.org/W4362655426","https://openalex.org/W4381245716","https://openalex.org/W4384918448","https://openalex.org/W4385572451"],"related_works":["https://openalex.org/W3197854638","https://openalex.org/W3140454661","https://openalex.org/W4245029315","https://openalex.org/W2139970489","https://openalex.org/W1492315459","https://openalex.org/W1512434910","https://openalex.org/W2540910169","https://openalex.org/W3089780453","https://openalex.org/W3148904318","https://openalex.org/W2022803902"],"abstract_inverted_index":{"Does":[0],"inverse":[1],"scaling":[2,109],"only":[3],"occur":[4,15],"as":[5],"a":[6,87,95],"function":[7],"of":[8,19,31,69,72,115],"model":[9],"size,":[10],"or":[11],"can":[12,37],"it":[13],"also":[14],"over":[16,66],"the":[17,29,47,67,100,113],"course":[18,68],"training?":[20],"We":[21,51],"carry":[22],"out":[23],"an":[24],"exploratory":[25],"study":[26],"investigating":[27],"whether":[28],"performance":[30,41,65,99,117,134],"language":[32,48,92],"models":[33,93,124],"on":[34,46,55,127],"specific":[35],"tasks":[36,54,74],"decrease":[38,97],"(while":[39],"general":[40],"remains":[42],"high)":[43],"during":[44],"training":[45],"modeling":[49],"task.":[50],"find":[52],"8":[53],"which":[56],"Pythia":[57],"12B":[58],"(Biderman":[59],"et":[60],"al.,":[61],"2023)":[62],"shows":[63],"decreased":[64],"training.":[70],"Five":[71],"these":[73],"(TruthfulQA-MC1,":[75],"TruthfulQA-MC2,":[76],"Hindsight":[77],"Neglect,":[78],"Memo":[79],"Trap,":[80],"and":[81],"Pattern":[82],"Match":[83],"Suppression)":[84],"additionally":[85],"show":[86,94],"consistent":[88],"relationship":[89],"whereby":[90],"larger":[91],"greater":[96],"in":[98],"more":[101],"they":[102],"are":[103,125],"trained,":[104],"despite":[105],"showing":[106],"standard":[107],"(positive)":[108],"overall.":[110],"This":[111],"highlights":[112],"importance":[114],"testing":[116],"at":[118],"all":[119],"relevant":[120],"benchmarks":[121],"any":[122],"time":[123],"trained":[126],"additional":[128],"data,":[129],"even":[130],"if":[131],"their":[132],"overall":[133],"improves.":[135]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-10T00:00:00"}