{"id":"https://openalex.org/W7164897106","doi":"https://doi.org/10.48550/arxiv.2606.16908","title":"LESS Is More: Mutual-Stability Sampling for Diffusion Language Models","display_name":"LESS Is More: Mutual-Stability Sampling for Diffusion Language Models","publication_year":2026,"publication_date":"2026-06-15","ids":{"openalex":"https://openalex.org/W7164897106","doi":"https://doi.org/10.48550/arxiv.2606.16908"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.16908","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.16908","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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.2606.16908","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5138737071","display_name":"Amr Mohamed","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mohamed, Amr","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5038162261","display_name":"Guokan Shang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shang, Guokan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5138746009","display_name":"Michalis Vazirgiannis","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Vazirgiannis, Michalis","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/T10201","display_name":"Speech Recognition and Synthesis","score":0.16009999811649323,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.16009999811649323,"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":0.14659999310970306,"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":0.10369999706745148,"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/security-token","display_name":"Security token","score":0.7305999994277954},{"id":"https://openalex.org/keywords/sampling","display_name":"Sampling (signal processing)","score":0.5407999753952026},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.5375000238418579},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.5160999894142151},{"id":"https://openalex.org/keywords/autoregressive-model","display_name":"Autoregressive model","score":0.5065000057220459},{"id":"https://openalex.org/keywords/importance-sampling","display_name":"Importance sampling","score":0.4862000048160553},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.48069998621940613},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.4747999906539917}],"concepts":[{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.7305999994277954},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6514000296592712},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.5407999753952026},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5375000238418579},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.5160999894142151},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5097000002861023},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.5065000057220459},{"id":"https://openalex.org/C52740198","wikidata":"https://www.wikidata.org/wiki/Q1539564","display_name":"Importance sampling","level":3,"score":0.4862000048160553},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.48069998621940613},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.4747999906539917},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.3959999978542328},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.39469999074935913},{"id":"https://openalex.org/C198082294","wikidata":"https://www.wikidata.org/wiki/Q3399648","display_name":"Position (finance)","level":2,"score":0.3492000102996826},{"id":"https://openalex.org/C61797465","wikidata":"https://www.wikidata.org/wiki/Q1188986","display_name":"Term (time)","level":2,"score":0.3240000009536743},{"id":"https://openalex.org/C163294075","wikidata":"https://www.wikidata.org/wiki/Q581861","display_name":"Noise reduction","level":2,"score":0.31769999861717224},{"id":"https://openalex.org/C167723999","wikidata":"https://www.wikidata.org/wiki/Q3773214","display_name":"Sampling distribution","level":2,"score":0.30979999899864197},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.3057999908924103},{"id":"https://openalex.org/C29265498","wikidata":"https://www.wikidata.org/wiki/Q7047719","display_name":"Noise measurement","level":3,"score":0.30309998989105225},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3012000024318695},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.265500009059906},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.25699999928474426}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.16908","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.16908","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.16908","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.16908","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"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":{"Diffusion":[0],"large":[1],"language":[2],"models":[3],"(dLLMs)":[4],"offer":[5],"a":[6,35,60,79,85,166],"promising":[7],"alternative":[8],"to":[9],"autoregressive":[10],"decoding":[11],"by":[12,30],"iteratively":[13],"refining":[14],"masked":[15,86],"sequences,":[16],"enabling":[17],"parallel":[18],"token":[19,67,101],"updates":[20],"and":[21,50,107,124,129,141,182],"bidirectional":[22],"conditioning.":[23],"Their":[24],"practical":[25],"efficiency,":[26],"however,":[27],"is":[28,111],"limited":[29],"sampling":[31,77,132],"procedures":[32],"that":[33,65,83],"execute":[34],"fixed":[36],"number":[37],"of":[38],"reverse":[39,105,156,163],"denoising":[40],"steps":[41,157],"selected":[42],"before":[43],"decoding,":[44],"spending":[45],"computation":[46],"on":[47,121],"already-stable":[48],"positions":[49],"sometimes":[51],"committing":[52],"unstable":[53],"ones":[54],"too":[55],"early.":[56],"We":[57,118],"present":[58],"\\textsc{LESS},":[59],"training-free,":[61],"model-agnostic":[62],"adaptive":[63,150],"sampler":[64],"treats":[66],"commitment":[68],"as":[69],"an":[70],"online":[71],"stopping":[72],"problem.":[73],"\\textsc{LESS}":[74,120,143],"implements":[75],"mutual-stability":[76],"through":[78],"joint":[80],"stability":[81],"rule":[82],"makes":[84],"position":[87],"eligible":[88],"for":[89],"unmasking":[90],"only":[91],"when":[92],"its":[93,99,108],"top-1":[94,100],"prediction":[95],"has":[96],"high":[97],"confidence,":[98],"persists":[102],"across":[103,134],"recent":[104],"steps,":[106],"predictive":[109],"distribution":[110],"stable":[112],"under":[113],"top-$K$":[114],"inter-step":[115],"Jensen--Shannon":[116],"divergence.":[117],"evaluate":[119],"Dream-7B,":[122],"LLaDA-8B,":[123],"LLaDA-1.5-8B,":[125],"covering":[126],"full-sequence":[127],"diffusion":[128],"semi-autoregressive":[130],"blockwise":[131],"regimes,":[133],"seven":[135],"benchmarks":[136],"spanning":[137],"general":[138],"knowledge,":[139],"math,":[140],"code.":[142],"improves":[144],"average":[145],"accuracy":[146],"over":[147],"strong":[148],"training-free":[149],"samplers":[151],"while":[152],"using":[153],"$72.1\\%$":[154],"fewer":[155,175],"than":[158],"fixed-budget":[159],"decoding.":[160],"Since":[161],"each":[162],"step":[164],"requires":[165],"Transformer":[167],"forward":[168,176],"pass,":[169],"these":[170],"step-count":[171],"reductions":[172],"translate":[173],"into":[174],"evaluations,":[177],"lower":[178,183],"measured":[179],"wall-clock":[180],"latency,":[181],"estimated":[184],"inference":[185],"compute.":[186]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-06-17T00:00:00"}