{"id":"https://openalex.org/W4416034151","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.1063","title":"On the Versatility of Sparse Autoencoders for In-Context Learning","display_name":"On the Versatility of Sparse Autoencoders for In-Context Learning","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4416034151","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.1063"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.1063","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.1063","pdf_url":"https://aclanthology.org/2025.findings-emnlp.1063.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 2025","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.findings-emnlp.1063.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5029401026","display_name":"Ikhyun Cho","orcid":"https://orcid.org/0009-0006-0360-256X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ikhyun Cho","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Gaeul Kwon","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gaeul Kwon","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5042661827","display_name":"Julia Hockenmaier","orcid":"https://orcid.org/0000-0003-2796-1752"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Julia Hockenmaier","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":3,"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":0.15468168,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"19531","last_page":"19538"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.2777000069618225,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.2777000069618225,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.22339999675750732,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T11448","display_name":"Face recognition and analysis","score":0.15700000524520874,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/feature","display_name":"Feature (linguistics)","score":0.37619999051094055},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.3138999938964844},{"id":"https://openalex.org/keywords/noise","display_name":"Noise (video)","score":0.2831999957561493},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.2741999924182892},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.24719999730587006}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6259999871253967},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6164000034332275},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.37619999051094055},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.3138999938964844},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2883000075817108},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.2831999957561493},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.2741999924182892},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.24719999730587006},{"id":"https://openalex.org/C59404180","wikidata":"https://www.wikidata.org/wiki/Q17013334","display_name":"Feature learning","level":2,"score":0.24320000410079956},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.24240000545978546}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-emnlp.1063","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.1063","pdf_url":"https://aclanthology.org/2025.findings-emnlp.1063.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 2025","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.1063","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.1063","pdf_url":"https://aclanthology.org/2025.findings-emnlp.1063.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 2025","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G4804575867","display_name":"Category I: Crossing the Divide Between Today's Practice and Tomorrow's Science","funder_award_id":"2005572","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G6953012893","display_name":null,"funder_award_id":"OAC 2005572","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320332180","display_name":"Defense Advanced Research Projects Agency","ror":"https://ror.org/02caytj08"},{"id":"https://openalex.org/F4320332222","display_name":"University of Illinois at Urbana-Champaign","ror":"https://ror.org/047426m28"},{"id":"https://openalex.org/F4320332815","display_name":"Advanced Research Projects Agency","ror":"https://ror.org/02caytj08"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416034151.pdf","grobid_xml":"https://content.openalex.org/works/W4416034151.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Sparse":[0],"autoencoders":[1],"(SAEs)":[2],"are":[3,47],"emerging":[4],"as":[5,61],"a":[6,65,105,110,118],"key":[7],"analytical":[8],"tool":[9],"in":[10,108],"the":[11,94,97,127],"field":[12],"of":[13,51,68,96,129],"mechanistic":[14],"interpretability":[15],"for":[16,26,53,122,135],"large":[17],"language":[18],"models":[19],"(LLMs).While":[20],"SAEs":[21,37,46,82,130],"have":[22],"primarily":[23],"been":[24],"used":[25],"interpretability,":[27],"we":[28,56,79,92],"shift":[29],"focus":[30],"and":[31,116,131],"explore":[32],"an":[33],"understudied":[34],"question:":[35],"\"Can":[36],"be":[38,84],"applied":[39,86],"to":[40,87],"practical":[41,74],"tasks":[42],"beyond":[43,138],"interpretability?\"Given":[44],"that":[45,71,81,102],"trained":[48],"on":[49,76],"billions":[50],"tokens":[52],"sparse":[54],"reconstruction,":[55],"believe":[57],"they":[58],"can":[59,72,83],"serve":[60],"effective":[62],"extractors,":[63],"offering":[64,117],"wide":[66],"range":[67],"useful":[69],"knowledge":[70],"benefit":[73],"applications.Building":[75],"this":[77],"motivation,":[78],"demonstrate":[80],"effectively":[85],"incontext":[88],"learning":[89],"(ICL).In":[90],"particular,":[91],"highlight":[93],"utility":[95],"SAE-reconstruction":[98],"loss":[99],"by":[100],"showing":[101],"it":[103],"provides":[104],"valuable":[106],"signal":[107],"ICL-exhibiting":[109],"strong":[111],"correlation":[112],"with":[113],"LLM":[114],"performance":[115],"powerful":[119],"unsupervised":[120],"approach":[121],"prompt":[123],"selection.These":[124],"findings":[125],"underscore":[126],"versatility":[128],"reveal":[132],"their":[133],"potential":[134],"real-world":[136],"applications":[137],"interpretability.Our":[139],"code":[140],"is":[141],"available":[142],"at":[143],"https://github.com/ihcho2/SAE-GPS.":[144]},"counts_by_year":[],"updated_date":"2026-06-22T08:00:12.763002","created_date":"2025-11-08T00:00:00"}