{"id":"https://openalex.org/W7125955842","doi":"https://doi.org/10.48550/arxiv.2601.19503","title":"GradPruner: Gradient-Guided Layer Pruning Enabling Efficient Fine-Tuning and Inference for LLMs","display_name":"GradPruner: Gradient-Guided Layer Pruning Enabling Efficient Fine-Tuning and Inference for LLMs","publication_year":2026,"publication_date":"2026-01-27","ids":{"openalex":"https://openalex.org/W7125955842","doi":"https://doi.org/10.48550/arxiv.2601.19503"},"language":null,"primary_location":{"id":"pmh:doi:10.48550/arxiv.2601.19503","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"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":"Article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":null,"any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5124115654","display_name":"Wei Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Wei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5038610148","display_name":"Anda Cheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cheng, Anda","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5124078694","display_name":"Yinggui Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yinggui","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":null,"has_fulltext":false,"cited_by_count":0,"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/T13702","display_name":"Machine Learning in Healthcare","score":0.38589999079704285,"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/T13702","display_name":"Machine Learning in Healthcare","score":0.38589999079704285,"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/T11636","display_name":"Artificial Intelligence in Healthcare and Education","score":0.29660001397132874,"subfield":{"id":"https://openalex.org/subfields/2718","display_name":"Health Informatics"},"field":{"id":"https://openalex.org/fields/27","display_name":"Medicine"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.10540000349283218,"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/inference","display_name":"Inference","score":0.757099986076355},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.6680999994277954},{"id":"https://openalex.org/keywords/merge","display_name":"Merge (version control)","score":0.6187999844551086},{"id":"https://openalex.org/keywords/downstream","display_name":"Downstream (manufacturing)","score":0.4537999927997589},{"id":"https://openalex.org/keywords/layer","display_name":"Layer (electronics)","score":0.4178999960422516},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.3961000144481659},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.35370001196861267}],"concepts":[{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.757099986076355},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6769000291824341},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.6680999994277954},{"id":"https://openalex.org/C197129107","wikidata":"https://www.wikidata.org/wiki/Q1921621","display_name":"Merge (version control)","level":2,"score":0.6187999844551086},{"id":"https://openalex.org/C2776207758","wikidata":"https://www.wikidata.org/wiki/Q5303302","display_name":"Downstream (manufacturing)","level":2,"score":0.4537999927997589},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4481000006198883},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.4244999885559082},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.4178999960422516},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.414900004863739},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.3961000144481659},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.35370001196861267},{"id":"https://openalex.org/C2780719617","wikidata":"https://www.wikidata.org/wiki/Q1030752","display_name":"Salient","level":2,"score":0.32429999113082886},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.3183000087738037},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.28870001435279846},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.2761000096797943},{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.27390000224113464},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2583000063896179},{"id":"https://openalex.org/C2776650193","wikidata":"https://www.wikidata.org/wiki/Q264661","display_name":"Obstacle","level":2,"score":0.25040000677108765}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:doi:10.48550/arxiv.2601.19503","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"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":"Article"},{"id":"doi:10.48550/arxiv.2601.19503","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2601.19503","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"pmh:doi:10.48550/arxiv.2601.19503","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"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":"Article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Fine-tuning":[0],"Large":[1],"Language":[2],"Models":[3],"(LLMs)":[4],"with":[5,126,132,174],"downstream":[6,61,153],"data":[7],"is":[8,183],"often":[9,30],"considered":[10],"time-consuming":[11],"and":[12,34,42,57,111,123,158],"expensive.":[13],"Structured":[14],"pruning":[15],"methods":[16],"are":[17,136],"primarily":[18],"employed":[19],"to":[20,50,96,105,138],"improve":[21],"the":[22,55,77,84,91,98,107,116,121,127,133],"inference":[23,58],"efficiency":[24,59],"of":[25,60,71,80,87,94,109,172],"pre-trained":[26],"models.":[27],"Meanwhile,":[28],"they":[29],"require":[31],"additional":[32],"time":[33],"memory":[35],"for":[36],"training,":[37],"knowledge":[38],"distillation,":[39],"structure":[40],"search,":[41],"other":[43],"strategies,":[44],"making":[45],"efficient":[46],"model":[47],"fine-tuning":[48,95],"challenging":[49],"achieve.":[51],"To":[52],"simultaneously":[53],"enhance":[54],"training":[56],"task":[62],"fine-tuning,":[63],"we":[64],"introduce":[65],"GradPruner,":[66],"which":[67],"can":[68],"prune":[69],"layers":[70,110,118],"LLMs":[72,150],"guided":[73],"by":[74],"gradients":[75,86],"in":[76,179],"early":[78],"stages":[79],"fine-tuning.":[81],"GradPruner":[82,166],"uses":[83],"cumulative":[85],"each":[88],"parameter":[89,170],"during":[90],"initial":[92],"phase":[93],"compute":[97],"Initial":[99],"Gradient":[100],"Information":[101],"Accumulation":[102],"Matrix":[103],"(IGIA-Matrix)":[104],"assess":[106],"importance":[108],"perform":[112],"pruning.":[113],"We":[114,144],"sparsify":[115],"pruned":[117],"based":[119],"on":[120,148],"IGIA-Matrix":[122],"merge":[124],"them":[125],"remaining":[128],"layers.":[129],"Only":[130],"elements":[131],"same":[134],"sign":[135,142],"merged":[137],"reduce":[139],"interference":[140],"from":[141],"variations.":[143],"conducted":[145],"extensive":[146],"experiments":[147],"two":[149],"across":[151],"eight":[152],"datasets.":[154],"Including":[155],"medical,":[156],"financial,":[157],"general":[159],"benchmark":[160],"tasks.":[161],"The":[162],"results":[163],"demonstrate":[164],"that":[165],"has":[167],"achieved":[168],"a":[169,176],"reduction":[171],"40%":[173],"only":[175],"0.99%":[177],"decrease":[178],"accuracy.":[180],"Our":[181],"code":[182],"publicly":[184],"available.":[185]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-01-29T00:00:00"}