{"id":"https://openalex.org/W7125864310","doi":"https://doi.org/10.48550/arxiv.2601.17869","title":"On the Emergence and Test-Time Use of Structural Information in Large Language Models","display_name":"On the Emergence and Test-Time Use of Structural Information in Large Language Models","publication_year":2026,"publication_date":"2026-01-25","ids":{"openalex":"https://openalex.org/W7125864310","doi":"https://doi.org/10.48550/arxiv.2601.17869"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2601.17869","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2601.17869","pdf_url":"https://arxiv.org/pdf/2601.17869","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2601.17869","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5124039311","display_name":"Michelle Chao Chen","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Michelle Chao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5003139809","display_name":"Moritz Miller","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Miller, Moritz","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5044005697","display_name":"Bernhard Sch\u00f6lkopf","orcid":"https://orcid.org/0000-0002-8177-0925","observed_orcids":["https://orcid.org/0000-0002-8177-0925"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sch\u00f6lkopf, Bernhard","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5124042697","display_name":"Siyuan Guo","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guo, Siyuan","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":0.8598999977111816,"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":0.8598999977111816,"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":0.10189999639987946,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.008200000040233135,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/structural-information","display_name":"structural information","score":0.7879999876022339},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.9279999732971191},{"id":"https://openalex.org/keywords/emergent-abilities","display_name":"emergent capabilities","score":0.5220000147819519},{"id":"https://openalex.org/keywords/mechanistic-understanding","display_name":"mechanistic understanding","score":0.675000011920929}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7032999992370605},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5300999879837036},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.5152999758720398},{"id":"https://openalex.org/C2779954242","wikidata":"https://www.wikidata.org/wiki/Q6031227","display_name":"Information structure","level":2,"score":0.5073000192642212},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.45590001344680786},{"id":"https://openalex.org/C2983448237","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Language understanding","level":2,"score":0.4336000084877014},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4113999903202057},{"id":"https://openalex.org/C59656382","wikidata":"https://www.wikidata.org/wiki/Q191536","display_name":"Conjunction (astronomy)","level":2,"score":0.39559999108314514},{"id":"https://openalex.org/C2776608160","wikidata":"https://www.wikidata.org/wiki/Q4785462","display_name":"Natural (archaeology)","level":2,"score":0.3822000026702881},{"id":"https://openalex.org/C2779439875","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Natural language understanding","level":3,"score":0.3255000114440918},{"id":"https://openalex.org/C188147891","wikidata":"https://www.wikidata.org/wiki/Q147638","display_name":"Cognitive science","level":1,"score":0.30660000443458557},{"id":"https://openalex.org/C195807954","wikidata":"https://www.wikidata.org/wiki/Q1662562","display_name":"Information extraction","level":2,"score":0.30660000443458557},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.30550000071525574},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.3052999973297119},{"id":"https://openalex.org/C94922259","wikidata":"https://www.wikidata.org/wiki/Q33215","display_name":"Constructed language","level":2,"score":0.2996000051498413},{"id":"https://openalex.org/C74672266","wikidata":"https://www.wikidata.org/wiki/Q815859","display_name":"Language acquisition","level":2,"score":0.28060001134872437},{"id":"https://openalex.org/C155092808","wikidata":"https://www.wikidata.org/wiki/Q182557","display_name":"Computational linguistics","level":2,"score":0.27250000834465027},{"id":"https://openalex.org/C66024118","wikidata":"https://www.wikidata.org/wiki/Q1122506","display_name":"Computational model","level":2,"score":0.2696000039577484},{"id":"https://openalex.org/C180198813","wikidata":"https://www.wikidata.org/wiki/Q121182","display_name":"Information system","level":2,"score":0.26269999146461487}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2601.17869","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2601.17869","pdf_url":"https://arxiv.org/pdf/2601.17869","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2601.17869","doi":"https://doi.org/10.48550/arxiv.2601.17869","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2601.17869","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,"listed_in":[],"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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2601.17869","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2601.17869","pdf_url":"https://arxiv.org/pdf/2601.17869","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4412887020","https://openalex.org/W4281488715","https://openalex.org/W4387560065","https://openalex.org/W7219092103","https://openalex.org/W4412944180","https://openalex.org/W7211955554","https://openalex.org/W7167818790","https://openalex.org/W4402670874","https://openalex.org/W4391871541","https://openalex.org/W7152568811"],"abstract_inverted_index":{"Learning":[0],"structural":[1,44,62,72],"information":[2,45,73],"from":[3],"observational":[4],"data":[5],"is":[6],"central":[7],"to":[8,83],"producing":[9],"new":[10],"knowledge":[11],"outside":[12],"the":[13,42,68,81],"training":[14],"corpus.":[15],"This":[16],"holds":[17],"for":[18],"mechanistic":[19],"understanding":[20],"in":[21],"scientific":[22],"discovery":[23],"as":[24,26],"well":[25],"flexible":[27],"test-time":[28,85],"compositional":[29,86],"generation.":[30],"We":[31,64],"thus":[32],"study":[33],"how":[34],"language":[35,57],"models":[36],"learn":[37],"abstract":[38],"structures":[39],"and":[40,79],"utilize":[41],"learnt":[43],"at":[46],"test-time.":[47],"To":[48],"ensure":[49],"a":[50,55],"controlled":[51],"setup,":[52],"we":[53],"design":[54],"natural":[56],"dataset":[58],"based":[59],"on":[60],"linguistic":[61],"transformations.":[63],"empirically":[65],"show":[66],"that":[67,80],"emergence":[69],"of":[70],"learning":[71],"correlates":[74],"with":[75],"complex":[76],"reasoning":[77],"tasks,":[78],"ability":[82],"perform":[84],"generation":[87],"remains":[88],"limited.":[89]},"counts_by_year":[],"updated_date":"2026-10-31T12:09:11.096373","created_date":"2026-01-28T00:00:00"}