{"id":"https://openalex.org/W7164568615","doi":"https://doi.org/10.48550/arxiv.2606.12841","title":"TimeROME-DLM: Temporal Causal Tracing and Low-Rank Inference-Time Knowledge Editing for Masked Diffusion Language Models","display_name":"TimeROME-DLM: Temporal Causal Tracing and Low-Rank Inference-Time Knowledge Editing for Masked Diffusion Language Models","publication_year":2026,"publication_date":"2026-06-11","ids":{"openalex":"https://openalex.org/W7164568615","doi":"https://doi.org/10.48550/arxiv.2606.12841"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.12841","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.12841","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2606.12841","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5126419543","display_name":"Zhengtao Yao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yao, Zhengtao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5037890950","display_name":"Liuyang Song","orcid":"https://orcid.org/0000-0003-4297-1668"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Song, Liuyang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138533798","display_name":"Hongbo Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Hongbo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5033109788","display_name":"Chenhao Wei","orcid":"https://orcid.org/0000-0002-2707-120X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wei, Chenhao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5110133782","display_name":"Haoyan Xu","orcid":"https://orcid.org/0009-0005-3103-8639"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Haoyan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138549855","display_name":"Guang Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Guang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5138508337","display_name":"Siheng Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Siheng","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":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/T10028","display_name":"Topic Modeling","score":0.4323999881744385,"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":0.4323999881744385,"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":0.04650000110268593,"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":0.03739999979734421,"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/tracing","display_name":"Tracing","score":0.4902999997138977},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.48969998955726624},{"id":"https://openalex.org/keywords/residual","display_name":"Residual","score":0.451200008392334},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.435699999332428},{"id":"https://openalex.org/keywords/generator","display_name":"Generator (circuit theory)","score":0.38199999928474426},{"id":"https://openalex.org/keywords/object","display_name":"Object (grammar)","score":0.3776000142097473},{"id":"https://openalex.org/keywords/autoregressive-model","display_name":"Autoregressive model","score":0.3594000041484833},{"id":"https://openalex.org/keywords/fraction","display_name":"Fraction (chemistry)","score":0.33000001311302185}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7416999936103821},{"id":"https://openalex.org/C138673069","wikidata":"https://www.wikidata.org/wiki/Q322229","display_name":"Tracing","level":2,"score":0.4902999997138977},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.48969998955726624},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.46399998664855957},{"id":"https://openalex.org/C155512373","wikidata":"https://www.wikidata.org/wiki/Q287450","display_name":"Residual","level":2,"score":0.451200008392334},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.435699999332428},{"id":"https://openalex.org/C2780992000","wikidata":"https://www.wikidata.org/wiki/Q17016113","display_name":"Generator (circuit theory)","level":3,"score":0.38199999928474426},{"id":"https://openalex.org/C2781238097","wikidata":"https://www.wikidata.org/wiki/Q175026","display_name":"Object (grammar)","level":2,"score":0.3776000142097473},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.3594000041484833},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3449999988079071},{"id":"https://openalex.org/C149629883","wikidata":"https://www.wikidata.org/wiki/Q660926","display_name":"Fraction (chemistry)","level":2,"score":0.33000001311302185},{"id":"https://openalex.org/C151201525","wikidata":"https://www.wikidata.org/wiki/Q177239","display_name":"Limit (mathematics)","level":2,"score":0.32519999146461487},{"id":"https://openalex.org/C2780719617","wikidata":"https://www.wikidata.org/wiki/Q1030752","display_name":"Salient","level":2,"score":0.31790000200271606},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.3059000074863434},{"id":"https://openalex.org/C2780385302","wikidata":"https://www.wikidata.org/wiki/Q367158","display_name":"Protocol (science)","level":3,"score":0.2922999858856201},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.28700000047683716},{"id":"https://openalex.org/C45357846","wikidata":"https://www.wikidata.org/wiki/Q2001982","display_name":"Notation","level":2,"score":0.2865999937057495},{"id":"https://openalex.org/C144745244","wikidata":"https://www.wikidata.org/wiki/Q4927286","display_name":"Blocking (statistics)","level":2,"score":0.2786000072956085},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.26840001344680786},{"id":"https://openalex.org/C34736171","wikidata":"https://www.wikidata.org/wiki/Q918333","display_name":"Preprocessor","level":2,"score":0.26499998569488525},{"id":"https://openalex.org/C126780896","wikidata":"https://www.wikidata.org/wiki/Q899871","display_name":"Distortion (music)","level":4,"score":0.257099986076355},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2549999952316284},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.2547000050544739},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.2513999938964844},{"id":"https://openalex.org/C23224414","wikidata":"https://www.wikidata.org/wiki/Q176769","display_name":"Hidden Markov model","level":2,"score":0.25130000710487366}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.12841","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.12841","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.12841","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.12841","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Masked":[0],"diffusion":[1,128],"language":[2],"models":[3],"(MDLMs)":[4],"such":[5],"as":[6],"LLaDA":[7],"now":[8],"rival":[9],"autoregressive":[10],"(AR)":[11],"LLMs,":[12],"but":[13],"every":[14,127],"existing":[15],"knowledge-editing":[16,65],"and":[17,26,49,99,110,117,176,214,228],"unlearning":[18],"method":[19],"(ROME,":[20],"MEMIT,":[21],"etc.)":[22],"targets":[23],"AR":[24,226],"transformers":[25],"either":[27],"makes":[28],"assumptions":[29],"that":[30,80,106,124],"fail":[31],"under":[32],"iterative":[33],"denoising,":[34],"or":[35],"requires":[36],"gradient":[37],"updates":[38],"whose":[39,87],"backward-pass":[40],"activations":[41],"cost":[42],"tens":[43],"of":[44,46,233],"GB":[45],"extra":[47],"VRAM":[48,207],"which":[50],"collapse":[51],"MDLMs":[52,229],"at":[53,95,123,126,187,230],"standard":[54],"learning":[55],"rates.":[56],"We":[57],"introduce":[58],"TimeROME-DLM,":[59],"the":[60,85,92,188,209,222,234],"first":[61],"training-free,":[62],"gradient-free,":[63],"inference-time":[64],"framework":[66],"for":[67,82],"MDLMs.":[68],"It":[69,178],"couples":[70],"two":[71],"components:":[72],"a":[73,100,119,149,198,231],"Temporal":[74],"Indirect":[75],"Effect":[76],"(TIE)":[77],"causal-tracing":[78],"protocol":[79],"identifies,":[81],"each":[83],"fact,":[84],"coordinate":[86,125],"intervention":[88],"most":[89],"strongly":[90],"drives":[91],"object":[93],"prediction":[94],"later":[96],"denoising":[97],"steps;":[98],"closed-form,":[101],"low-rank":[102],"residual":[103],"edit":[104],"memory":[105],"aggregates":[107],"subject":[108],"keys":[109],"target":[111],"deltas":[112],"across":[113,192],"all":[114],"forget":[115],"facts":[116],"applies":[118],"single":[120],"ridge-regularised":[121],"update":[122],"forward,":[129],"with":[130,156,204],"sparsification":[131],"to":[132,171,200,217],"limit":[133],"utility":[134],"spillover.":[135],"Backbone":[136],"weights":[137],"stay":[138],"frozen;":[139],"only":[140],"three":[141],"hyperparameters":[142],"(alpha,":[143],"lambda,":[144],"q)":[145],"are":[146],"tuned":[147],"on":[148],"small":[150],"validation":[151],"split.":[152],"On":[153],"TOFU":[154],"forget01":[155],"TOFU-finetuned":[157],"LLaDA-8B-Base,":[158],"TimeROME-DLM":[159,220],"cuts":[160],"forget-set":[161],"log-probability":[162,181],"by":[163],"roughly":[164],"83":[165],"nats.":[166],"The":[167],"same":[168],"configuration":[169],"transfers":[170],"LLaDA-8B-Instruct,":[172],"Dream-7B,":[173],"MMaDA-8B,":[174],"DiffuLLaMA-7B,":[175],"LLaDA-MoE-1.4B.":[177],"keeps":[179],"retain-set":[180],"nearly":[182],"flat":[183],"(within":[184],"~1":[185],"nat":[186],"utility-safe":[189],"operating":[190],"point)":[191],"50":[193],"sequentially":[194],"inserted":[195],"facts,":[196],"delivers":[197],"four-":[199],"fourteen-fold":[201],"wall-clock":[202],"speedup":[203],"zero":[205],"additional":[206],"over":[208],"strongest":[210],"converged":[211],"training-time":[212],"baseline,":[213],"scales":[215],"sub-linearly":[216],"400":[218],"facts.":[219],"closes":[221],"locate-then-edit":[223],"gap":[224],"between":[225],"LLMs":[227],"fraction":[232],"computational":[235],"cost.":[236]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-06-13T00:00:00"}