{"id":"https://openalex.org/W7163192299","doi":"https://doi.org/10.48550/arxiv.2606.00148","title":"StemBind: When MLLMs Get Lost Between Rules and Instances in Abstract Visual Reasoning","display_name":"StemBind: When MLLMs Get Lost Between Rules and Instances in Abstract Visual Reasoning","publication_year":2026,"publication_date":"2026-05-29","ids":{"openalex":"https://openalex.org/W7163192299","doi":"https://doi.org/10.48550/arxiv.2606.00148"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.00148","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.00148","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":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.2606.00148","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5041911045","display_name":"Xixiang He","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"He, Xixiang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137710323","display_name":"Baiqi Wu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Baiqi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5058492670","display_name":"Xingming Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Xingming","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130187759","display_name":"Ao Cheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cheng, Ao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137679754","display_name":"Qiyao Sun","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sun, Qiyao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130004621","display_name":"Xuanyu Ji","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ji, Xuanyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5036921959","display_name":"Qingyong Hu","orcid":"https://orcid.org/0000-0003-0337-9207"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hu, Qingyong","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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9294999837875366,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9294999837875366,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.014100000262260437,"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/T10028","display_name":"Topic Modeling","score":0.00570000009611249,"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/visual-reasoning","display_name":"Visual reasoning","score":0.583299994468689},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.5667999982833862},{"id":"https://openalex.org/keywords/perception","display_name":"Perception","score":0.40560001134872437},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.40529999136924744},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.3962000012397766},{"id":"https://openalex.org/keywords/grammaticality","display_name":"Grammaticality","score":0.36640000343322754},{"id":"https://openalex.org/keywords/rule-based-system","display_name":"Rule-based system","score":0.36079999804496765},{"id":"https://openalex.org/keywords/analytic-reasoning","display_name":"Analytic reasoning","score":0.335099995136261},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.32899999618530273}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6741999983787537},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6391000151634216},{"id":"https://openalex.org/C2777508537","wikidata":"https://www.wikidata.org/wiki/Q7936620","display_name":"Visual reasoning","level":2,"score":0.583299994468689},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.5667999982833862},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.5088000297546387},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4194999933242798},{"id":"https://openalex.org/C26760741","wikidata":"https://www.wikidata.org/wiki/Q160402","display_name":"Perception","level":2,"score":0.40560001134872437},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.40529999136924744},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.3962000012397766},{"id":"https://openalex.org/C2779525943","wikidata":"https://www.wikidata.org/wiki/Q1187300","display_name":"Grammaticality","level":3,"score":0.36640000343322754},{"id":"https://openalex.org/C149271511","wikidata":"https://www.wikidata.org/wiki/Q1417149","display_name":"Rule-based system","level":2,"score":0.36079999804496765},{"id":"https://openalex.org/C103057564","wikidata":"https://www.wikidata.org/wiki/Q4751139","display_name":"Analytic reasoning","level":3,"score":0.335099995136261},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.32899999618530273},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.32019999623298645},{"id":"https://openalex.org/C178253425","wikidata":"https://www.wikidata.org/wiki/Q162668","display_name":"Visual perception","level":3,"score":0.3009999990463257},{"id":"https://openalex.org/C2779918689","wikidata":"https://www.wikidata.org/wiki/Q3771842","display_name":"Stimulus (psychology)","level":2,"score":0.29409998655319214},{"id":"https://openalex.org/C184898388","wikidata":"https://www.wikidata.org/wiki/Q1435712","display_name":"Pairwise comparison","level":2,"score":0.2856999933719635},{"id":"https://openalex.org/C189430467","wikidata":"https://www.wikidata.org/wiki/Q7293293","display_name":"Ranking (information retrieval)","level":2,"score":0.28060001134872437},{"id":"https://openalex.org/C552651612","wikidata":"https://www.wikidata.org/wiki/Q7328862","display_name":"Visual thinking","level":2,"score":0.2770000100135803},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.2732999920845032},{"id":"https://openalex.org/C49937458","wikidata":"https://www.wikidata.org/wiki/Q2599292","display_name":"Probabilistic logic","level":2,"score":0.27000001072883606},{"id":"https://openalex.org/C37335422","wikidata":"https://www.wikidata.org/wiki/Q6888134","display_name":"Model-based reasoning","level":3,"score":0.2680000066757202},{"id":"https://openalex.org/C205711294","wikidata":"https://www.wikidata.org/wiki/Q176953","display_name":"Rendering (computer graphics)","level":2,"score":0.26249998807907104},{"id":"https://openalex.org/C2780103172","wikidata":"https://www.wikidata.org/wiki/Q1309721","display_name":"Visual Objects","level":3,"score":0.26249998807907104},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.2556999921798706}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.00148","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.00148","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":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.00148","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.00148","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":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"display_name":"Peace, Justice and strong institutions","score":0.6341552138328552,"id":"https://metadata.un.org/sdg/16"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Multimodal":[0],"large":[1],"language":[2],"models":[3,190,226],"(MLLMs)":[4],"often":[5],"know":[6],"the":[7,11,29,37,69,81,106,169,187,197,211,233],"rule":[8,50,154,170,239],"but":[9],"pick":[10],"wrong":[12],"answer:":[13],"on":[14,105,159,186],"abstract":[15,253],"visual":[16,71,118,254],"reasoning":[17,132,255],"(AVR)":[18],"tasks,":[19,123],"a":[20,56,63,95,102,262],"model":[21],"can":[22,98],"describe":[23],"what":[24],"it":[25],"sees":[26],"and":[27,52,88,181,206,219,235,240],"name":[28],"underlying":[30],"pattern,":[31],"yet":[32],"still":[33,191],"fail":[34],"to":[35,101,214,250],"choose":[36],"matching":[38],"candidate.":[39],"Existing":[40],"AVR":[41,245],"benchmarks":[42],"cannot":[43],"detect":[44],"this":[45],"because":[46],"they":[47],"collapse":[48],"perception,":[49],"induction,":[51],"answer":[53,192],"selection":[54],"into":[55],"single":[57],"right-or-wrong":[58],"signal.":[59],"We":[60],"introduce":[61],"StemBind,":[62],"shared-stem":[64],"diagnostic":[65],"benchmark":[66],"that":[67],"probes":[68],"same":[70,107,188],"stem":[72],"with":[73,124],"three":[74],"aligned":[75],"questions:":[76],"Perception":[77],"(what":[78,84],"is":[79,171,202],"in":[80],"image),":[82],"Rule":[83],"pattern":[85],"governs":[86],"it),":[87,93],"Full":[89],"(which":[90],"option":[91],"completes":[92],"so":[94,164],"final-answer":[96,248],"error":[97],"be":[99],"attributed":[100],"specific":[103],"sub-step":[104],"evidence.":[108],"StemBind":[109,243],"contains":[110],"2,298":[111],"curated":[112],"knowledge-light":[113],"stems":[114],"across":[115],"nine":[116],"auditable":[117],"operations,":[119],"totaling":[120],"19,533":[121],"P/R/F":[122],"each":[125],"full":[126],"item":[127],"annotated":[128],"by":[129],"Sternberg's":[130],"four":[131,148],"stages":[133],"(S1":[134],"Encode,":[135],"S2":[136],"Infer,":[137],"S3":[138],"Map,":[139],"S4":[140],"Apply).":[141],"Evaluating":[142],"24":[143,162],"frontier":[144],"MLLM":[145],"configurations":[146],"yields":[147],"findings.":[149],"(i)":[150],"The":[151,200],"R-F":[152],"chasm:":[153],"accuracy":[155,158],"exceeds":[156],"full-item":[157,241],"22":[160],"of":[161,196],"models,":[163],"most":[165],"failures":[166],"happen":[167],"after":[168],"identified.":[172],"(ii)":[173],"A":[174],"persistent":[175],"binding":[176,260],"gap:":[177],"even":[178,237],"when":[179],"P":[180],"R":[182],"are":[183],"both":[184],"correct":[185],"stem,":[189],"F":[193],"incorrectly":[194],"51.2%":[195],"time.":[198],"(iii)":[199],"bottleneck":[201],"S3:":[203],"process":[204],"diagnostics":[205],"Stage-wise":[207],"Stimulus":[208],"Augmentation":[209],"localize":[210],"dominant":[212],"failure":[213],"rule-to-instance":[215,259],"mapping.":[216],"(iv)":[217],"Scaling":[218],"thinking":[220,229,236],"do":[221],"not":[222],"help:":[223],"neither":[224],"larger":[225],"nor":[227],"explicit":[228],"mode":[230],"reliably":[231],"closes":[232],"gap,":[234],"lowers":[238],"accuracy.":[242],"reframes":[244],"evaluation":[246],"from":[247],"ranking":[249],"locating":[251],"where":[252],"breaks":[256],"down,":[257],"identifying":[258],"as":[261],"concrete":[263],"next":[264],"target":[265],"for":[266],"vision-grounded":[267],"reasoning.":[268]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-03T00:00:00"}