{"id":"https://openalex.org/W4412888021","doi":"https://doi.org/10.18653/v1/2025.findings-acl.768","title":"Improving Efficiency in Large Language Models via Extendable Block Floating Point Representation","display_name":"Improving Efficiency in Large Language Models via Extendable Block Floating Point Representation","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412888021","doi":"https://doi.org/10.18653/v1/2025.findings-acl.768"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.findings-acl.768","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.768","pdf_url":"https://aclanthology.org/2025.findings-acl.768.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: ACL 2025","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.findings-acl.768.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100459316","display_name":"Dongyang Li","orcid":"https://orcid.org/0000-0001-5513-360X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dongyang Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100779776","display_name":"Zeyang Li","orcid":"https://orcid.org/0000-0001-6841-5905"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zeyang Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5068283310","display_name":"Bosheng Liu","orcid":"https://orcid.org/0000-0001-7123-7425"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bosheng Liu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5060755096","display_name":"Jigang Wu","orcid":"https://orcid.org/0000-0002-6470-9794"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jigang Wu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"14861","last_page":"14873"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9860000014305115,"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":0.9860000014305115,"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":0.9851999878883362,"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/computer-science","display_name":"Computer science","score":0.7087979316711426},{"id":"https://openalex.org/keywords/block","display_name":"Block (permutation group theory)","score":0.625478982925415},{"id":"https://openalex.org/keywords/representation","display_name":"Representation (politics)","score":0.6075035333633423},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.5595979690551758},{"id":"https://openalex.org/keywords/point","display_name":"Point (geometry)","score":0.5238380432128906},{"id":"https://openalex.org/keywords/intermediate-language","display_name":"Intermediate language","score":0.49181678891181946},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.4571113884449005},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.14836901426315308},{"id":"https://openalex.org/keywords/compiler","display_name":"Compiler","score":0.08795160055160522}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7087979316711426},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.625478982925415},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.6075035333633423},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.5595979690551758},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.5238380432128906},{"id":"https://openalex.org/C77660490","wikidata":"https://www.wikidata.org/wiki/Q244916","display_name":"Intermediate language","level":3,"score":0.49181678891181946},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.4571113884449005},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.14836901426315308},{"id":"https://openalex.org/C169590947","wikidata":"https://www.wikidata.org/wiki/Q47506","display_name":"Compiler","level":2,"score":0.08795160055160522},{"id":"https://openalex.org/C94625758","wikidata":"https://www.wikidata.org/wiki/Q7163","display_name":"Politics","level":2,"score":0.0},{"id":"https://openalex.org/C199539241","wikidata":"https://www.wikidata.org/wiki/Q7748","display_name":"Law","level":1,"score":0.0},{"id":"https://openalex.org/C17744445","wikidata":"https://www.wikidata.org/wiki/Q36442","display_name":"Political science","level":0,"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/v1/2025.findings-acl.768","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.768","pdf_url":"https://aclanthology.org/2025.findings-acl.768.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: ACL 2025","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-acl.768","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.768","pdf_url":"https://aclanthology.org/2025.findings-acl.768.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: ACL 2025","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G1409336737","display_name":null,"funder_award_id":"62302102","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G7783709206","display_name":null,"funder_award_id":"2023A1515012844","funder_id":"https://openalex.org/F4320337111","funder_display_name":"Basic and Applied Basic Research Foundation of Guangdong Province"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320337111","display_name":"Basic and Applied Basic Research Foundation of Guangdong Province","ror":null}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412888021.pdf","grobid_xml":"https://content.openalex.org/works/W4412888021.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W2062195135","https://openalex.org/W2372403409","https://openalex.org/W4242726756","https://openalex.org/W2029210135","https://openalex.org/W2932089593","https://openalex.org/W2064538205","https://openalex.org/W2469514515","https://openalex.org/W1483168212","https://openalex.org/W1591838558","https://openalex.org/W2571742053"],"abstract_inverted_index":{"Large":[0],"language":[1,7],"models":[2],"(LLMs)":[3],"have":[4],"revolutionized":[5],"natural":[6],"processing":[8],"(NLP)":[9],"tasks,":[10],"yet":[11],"their":[12],"increasing":[13,106],"size":[14],"poses":[15],"substantial":[16],"challenges":[17],"in":[18,45,148],"terms":[19],"of":[20,36,136],"computational":[21,49,152],"and":[22,39,48,122,151],"memory":[23],"resources.Block":[24],"floating-point":[25,38],"(BFP)":[26],"arithmetic":[27],"offers":[28],"an":[29],"effective":[30],"solution":[31],"by":[32,94],"leveraging":[33],"the":[34,82,100,107,118,134],"strengths":[35],"both":[37,46,149],"fixed-point":[40],"representations,":[41],"leading":[42,62],"to":[43,58,63,86,132],"reductions":[44],"storage":[47,115],"overhead.However,":[50],"current":[51],"low-bit":[52],"BFP":[53,78],"quantization":[54],"approaches":[55],"often":[56],"struggle":[57],"handle":[59],"extreme":[60],"outliers,":[61],"significant":[64,102],"accuracy":[65,121,150],"degradation.To":[66],"overcome":[67],"this":[68,93],"limitation,":[69],"we":[70],"introduce":[71],"Extendable":[72],"Exponent":[73],"Sharing":[74],"(EES),":[75],"a":[76,88,127],"novel":[77],"representation":[79],"that":[80,143],"extends":[81],"exponent":[83,97,138],"bit":[84,110,139],"width":[85,111],"capture":[87],"wider":[89],"dynamic":[90],"range.EES":[91],"achieves":[92],"embedding":[95],"extendable":[96,137],"bits":[98],"into":[99],"least":[101],"mantissa":[103],"bits,":[104],"thereby":[105],"shared":[108],"exponent's":[109],"without":[112],"incurring":[113],"additional":[114],"costs.To":[116],"optimize":[117,133],"trade-off":[119],"between":[120],"energy":[123],"efficiency,":[124],"EES":[125,144],"employs":[126],"design":[128],"space":[129],"exploration":[130],"strategy":[131],"configuration":[135],"widths.Experimental":[140],"results":[141],"show":[142],"outperforms":[145],"representative":[146],"baselines":[147],"efficiency.":[153]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}