{"id":"https://openalex.org/W4387939340","doi":"https://doi.org/10.20944/preprints202310.1487.v1","title":"An Efficient Plug-and-Play Post-Training Pruning Strategy in Large Language Models","display_name":"An Efficient Plug-and-Play Post-Training Pruning Strategy in Large Language Models","publication_year":2023,"publication_date":"2023-10-24","ids":{"openalex":"https://openalex.org/W4387939340","doi":"https://doi.org/10.20944/preprints202310.1487.v1"},"language":"en","primary_location":{"id":"doi:10.20944/preprints202310.1487.v1","doi":"https://doi.org/10.20944/preprints202310.1487.v1","is_oa":true,"landing_page_url":"https://doi.org/10.20944/preprints202310.1487.v1","pdf_url":"https://www.preprints.org/manuscript/202310.1487/v1/download","source":{"id":"https://openalex.org/S6309402219","display_name":"Preprints.org","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":["https://openalex.org/P4310310987"],"host_organization_lineage_names":["Multidisciplinary Digital Publishing Institute"],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"posted-content"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://www.preprints.org/manuscript/202310.1487/v1/download","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5032429023","display_name":"Yingtao Zhang","orcid":"https://orcid.org/0000-0002-8587-9645","observed_orcids":["https://orcid.org/0000-0002-8587-9645"]},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Yingtao Zhang","raw_affiliation_strings":["Department of Computer Science, Tsinghua University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science, Tsinghua University","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5080943673","display_name":"Haoli Bai","orcid":"https://orcid.org/0000-0002-2436-4875","observed_orcids":["https://orcid.org/0000-0002-2436-4875"]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Haoli Bai","raw_affiliation_strings":["Huawei Noah's Ark Lab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Noah's Ark Lab","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5025724191","display_name":"Haokun Lin","orcid":"https://orcid.org/0009-0000-1084-7115","observed_orcids":["https://orcid.org/0009-0000-1084-7115","https://orcid.org/0009-0003-8776-7780","https://orcid.org/0009-0009-9116-0432","https://orcid.org/0009-0004-8395-392X"]},"institutions":[{"id":"https://openalex.org/I19820366","display_name":"Chinese Academy of Sciences","ror":"https://ror.org/034t30j35","country_code":"CN","type":"government","lineage":["https://openalex.org/I19820366"]},{"id":"https://openalex.org/I4210112150","display_name":"Institute of Automation","ror":"https://ror.org/022c3hy66","country_code":"CN","type":"facility","lineage":["https://openalex.org/I19820366","https://openalex.org/I4210112150"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Haokun Lin","raw_affiliation_strings":["Institute of Automation, Chinese Academy of Sciences"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Institute of Automation, Chinese Academy of Sciences","institution_ids":["https://openalex.org/I19820366","https://openalex.org/I4210112150"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5009598811","display_name":"Jialin Zhao","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jialin Zhao","raw_affiliation_strings":["Department of Computer Science, Tsinghua University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science, Tsinghua University","institution_ids":["https://openalex.org/I99065089"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101663633","display_name":"Lu Hou","orcid":"https://orcid.org/0000-0002-4694-1821","observed_orcids":["https://orcid.org/0000-0002-4694-1821","https://orcid.org/0000-0003-3085-9353"]},"institutions":[{"id":"https://openalex.org/I2250955327","display_name":"Huawei Technologies (China)","ror":"https://ror.org/00cmhce21","country_code":"CN","type":"company","lineage":["https://openalex.org/I2250955327"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Lu Hou","raw_affiliation_strings":["Huawei Noah's Ark Lab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Noah's Ark Lab","institution_ids":["https://openalex.org/I2250955327"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5007730744","display_name":"Carlo Vittorio Cannistraci","orcid":"https://orcid.org/0000-0003-0100-8410","observed_orcids":["https://orcid.org/0000-0003-0100-8410"]},"institutions":[{"id":"https://openalex.org/I99065089","display_name":"Tsinghua University","ror":"https://ror.org/03cve4549","country_code":"CN","type":"education","lineage":["https://openalex.org/I99065089"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Carlo Vittorio Cannistraci","raw_affiliation_strings":["Department of Computer Science, Tsinghua University"],"raw_orcid":"https://orcid.org/0000-0003-0100-8410","affiliations":[{"raw_affiliation_string":"Department of Computer Science, Tsinghua University","institution_ids":["https://openalex.org/I99065089"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5007730744","https://openalex.org/A5032429023"],"corresponding_institution_ids":["https://openalex.org/I99065089"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":4,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.9688000082969666,"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"}},"topics":[{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9688000082969666,"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/T10028","display_name":"Topic Modeling","score":0.018699999898672104,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.0017999999690800905,"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/post-training-pruning","display_name":"post-training pruning","score":0.9729999899864197},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.9950000047683716},{"id":"https://openalex.org/keywords/model-compression","display_name":"model compression","score":0.7289999723434448},{"id":"https://openalex.org/keywords/llama","display_name":"llamas","score":0.7940000295639038}],"concepts":[{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.8627469539642334},{"id":"https://openalex.org/C21308566","wikidata":"https://www.wikidata.org/wiki/Q7169365","display_name":"Permutation (music)","level":2,"score":0.6840177774429321},{"id":"https://openalex.org/C176217482","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Metric (unit)","level":2,"score":0.570747971534729},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5465764999389648},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.540831983089447},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5246313214302063},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4523915648460388},{"id":"https://openalex.org/C127162648","wikidata":"https://www.wikidata.org/wiki/Q16858953","display_name":"Channel (broadcasting)","level":2,"score":0.4133087992668152},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.28942346572875977},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2808440029621124},{"id":"https://openalex.org/C21547014","wikidata":"https://www.wikidata.org/wiki/Q1423657","display_name":"Operations management","level":1,"score":0.193484365940094},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.158320814371109},{"id":"https://openalex.org/C205649164","wikidata":"https://www.wikidata.org/wiki/Q1071","display_name":"Geography","level":0,"score":0.1179283857345581},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.10121670365333557},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.0},{"id":"https://openalex.org/C153294291","wikidata":"https://www.wikidata.org/wiki/Q25261","display_name":"Meteorology","level":1,"score":0.0},{"id":"https://openalex.org/C24890656","wikidata":"https://www.wikidata.org/wiki/Q82811","display_name":"Acoustics","level":1,"score":0.0},{"id":"https://openalex.org/C6557445","wikidata":"https://www.wikidata.org/wiki/Q173113","display_name":"Agronomy","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.20944/preprints202310.1487.v1","doi":"https://doi.org/10.20944/preprints202310.1487.v1","is_oa":true,"landing_page_url":"https://doi.org/10.20944/preprints202310.1487.v1","pdf_url":"https://www.preprints.org/manuscript/202310.1487/v1/download","source":{"id":"https://openalex.org/S6309402219","display_name":"Preprints.org","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":["https://openalex.org/P4310310987"],"host_organization_lineage_names":["Multidisciplinary Digital Publishing Institute"],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"posted-content"}],"best_oa_location":{"id":"doi:10.20944/preprints202310.1487.v1","doi":"https://doi.org/10.20944/preprints202310.1487.v1","is_oa":true,"landing_page_url":"https://doi.org/10.20944/preprints202310.1487.v1","pdf_url":"https://www.preprints.org/manuscript/202310.1487/v1/download","source":{"id":"https://openalex.org/S6309402219","display_name":"Preprints.org","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":["https://openalex.org/P4310310987"],"host_organization_lineage_names":["Multidisciplinary Digital Publishing Institute"],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"posted-content"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4387939340.pdf","grobid_xml":"https://content.openalex.org/works/W4387939340.grobid-xml"},"referenced_works_count":32,"referenced_works":["https://openalex.org/W1632114991","https://openalex.org/W2156150815","https://openalex.org/W2222512263","https://openalex.org/W2525332836","https://openalex.org/W2764043458","https://openalex.org/W2808133870","https://openalex.org/W2894740066","https://openalex.org/W2896457183","https://openalex.org/W2954698171","https://openalex.org/W2981852735","https://openalex.org/W3024171804","https://openalex.org/W3035180000","https://openalex.org/W3127067080","https://openalex.org/W3131107792","https://openalex.org/W3132669550","https://openalex.org/W3156528192","https://openalex.org/W3175195442","https://openalex.org/W3211718241","https://openalex.org/W3212680907","https://openalex.org/W4226078967","https://openalex.org/W4229005866","https://openalex.org/W4292692470","https://openalex.org/W4307934016","https://openalex.org/W4309591680","https://openalex.org/W4313484599","https://openalex.org/W4322718191","https://openalex.org/W4379260375","https://openalex.org/W4380346028","https://openalex.org/W4381586827","https://openalex.org/W4384918448","https://openalex.org/W6893809404","https://openalex.org/W7133210858"],"related_works":["https://openalex.org/W4399051247","https://openalex.org/W7077456798","https://openalex.org/W7161739926","https://openalex.org/W7128542154","https://openalex.org/W7118186516","https://openalex.org/W4411811268","https://openalex.org/W4415795942","https://openalex.org/W4391899168","https://openalex.org/W7171568314","https://openalex.org/W4381586827"],"abstract_inverted_index":{"With":[0],"the":[1,29,35,73,106,146],"rapid":[2],"growth":[3],"of":[4,24,51],"large":[5],"language":[6],"models":[7],"(LLMs),":[8],"there":[9],"is":[10,37],"increasing":[11],"demand":[12],"for":[13,17,48],"memory":[14],"and":[15,32,63,75,80],"computation":[16],"LLMs.":[18,52,110],"Recent":[19],"efforts":[20],"on":[21,78,126,150,157],"post-training":[22,49,123],"pruning":[23,50,68,124,139],"LLMs":[25],"aim":[26],"to":[27,87,103,134],"reduce":[28],"model":[30],"size":[31],"computation,":[33],"yet":[34],"performance":[36],"still":[38],"sub-optimal.":[39],"In":[40],"this":[41],"paper,":[42],"we":[43],"present":[44],"a":[45,66,84],"plug-and-play":[46],"solution":[47,55],"The":[53,95],"proposed":[54,96],"has":[56],"two":[57,97],"innovative":[58],"components:":[59],"1)":[60],"**Relative":[61],"Importance":[62],"Activations**":[64],"(RIA),":[65],"new":[67,85],"metric":[69],"that":[70,115],"jointly":[71],"considers":[72],"weight":[74],"activations":[76],"efficiently":[77],"LLMs;":[79],"2)":[81],"**Channel":[82],"Permutation**,":[83],"approach":[86],"maximally":[88],"preserve":[89],"important":[90],"weights":[91],"under":[92],"N:M":[93,107,137],"sparsity.":[94],"components":[98],"can":[99,118,143],"be":[100],"readily":[101],"combined":[102],"further":[104],"enhance":[105],"structuredly":[108],"pruned":[109],"Our":[111],"empirical":[112],"experiments":[113],"show":[114],"RIA":[116],"alone":[117],"already":[119],"surpass":[120],"all":[121],"existing":[122],"methods":[125],"prevalent":[127],"LLMs,":[128],"e.g.,":[129],"LLaMA":[130],"ranging":[131],"from":[132],"7B":[133],"65B.":[135],"Furthermore,":[136],"structured":[138],"with":[140,154],"channel":[141],"permutation":[142],"even":[144],"outperform":[145],"original":[147],"LLaMA2":[148],"70B":[149],"zero-shot":[151],"tasks,":[152],"together":[153],"practical":[155],"speed-up":[156],"specific":[158],"hardware.":[159]},"counts_by_year":[{"year":2025,"cited_by_count":1},{"year":2024,"cited_by_count":2},{"year":2023,"cited_by_count":1}],"updated_date":"2026-11-09T13:22:03.289127","created_date":"2023-10-26T00:00:00"}