{"id":"https://openalex.org/W7152556640","doi":"https://doi.org/10.48550/arxiv.2604.06427","title":"The Depth Ceiling: On the Limits of Large Language Models in Discovering Latent Planning","display_name":"The Depth Ceiling: On the Limits of Large Language Models in Discovering Latent Planning","publication_year":2026,"publication_date":"2026-04-07","ids":{"openalex":"https://openalex.org/W7152556640","doi":"https://doi.org/10.48550/arxiv.2604.06427"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.06427","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.06427","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":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.2604.06427","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5035425640","display_name":"Yi Xu","orcid":"https://orcid.org/0000-0002-0604-8481"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Yi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5093126713","display_name":"Philipp Jettkant","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jettkant, Philipp","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5060010430","display_name":"Laura Ruis","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ruis, Laura","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/T10906","display_name":"AI-based Problem Solving and Planning","score":0.35690000653266907,"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/T10906","display_name":"AI-based Problem Solving and Planning","score":0.35690000653266907,"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":0.09830000251531601,"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":0.08219999819993973,"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/latent-class-model","display_name":"Latent class model","score":0.5232999920845032},{"id":"https://openalex.org/keywords/credence","display_name":"Credence","score":0.4961000084877014},{"id":"https://openalex.org/keywords/probabilistic-latent-semantic-analysis","display_name":"Probabilistic latent semantic analysis","score":0.4212999939918518},{"id":"https://openalex.org/keywords/latent-variable","display_name":"Latent variable","score":0.3538999855518341},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.32429999113082886},{"id":"https://openalex.org/keywords/latent-semantic-analysis","display_name":"Latent semantic analysis","score":0.3190000057220459}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6302000284194946},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5723000168800354},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5286999940872192},{"id":"https://openalex.org/C70727504","wikidata":"https://www.wikidata.org/wiki/Q1806878","display_name":"Latent class model","level":2,"score":0.5232999920845032},{"id":"https://openalex.org/C2779513410","wikidata":"https://www.wikidata.org/wiki/Q25351567","display_name":"Credence","level":2,"score":0.4961000084877014},{"id":"https://openalex.org/C112933361","wikidata":"https://www.wikidata.org/wiki/Q2845258","display_name":"Probabilistic latent semantic analysis","level":2,"score":0.4212999939918518},{"id":"https://openalex.org/C51167844","wikidata":"https://www.wikidata.org/wiki/Q4422623","display_name":"Latent variable","level":2,"score":0.3538999855518341},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.32429999113082886},{"id":"https://openalex.org/C170133592","wikidata":"https://www.wikidata.org/wiki/Q1806883","display_name":"Latent semantic analysis","level":2,"score":0.3190000057220459},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.31470000743865967},{"id":"https://openalex.org/C86251818","wikidata":"https://www.wikidata.org/wiki/Q816754","display_name":"Benchmarking","level":2,"score":0.29420000314712524},{"id":"https://openalex.org/C114073186","wikidata":"https://www.wikidata.org/wiki/Q2631895","display_name":"Automated planning and scheduling","level":2,"score":0.2621000111103058}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.06427","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.06427","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":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.2604.06427","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.06427","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"score":0.41089576482772827,"display_name":"Sustainable cities and communities","id":"https://metadata.un.org/sdg/11"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"The":[0],"viability":[1],"of":[2,25,67],"chain-of-thought":[3],"(CoT)":[4],"monitoring":[5],"hinges":[6],"on":[7,46],"models":[8,38,113],"being":[9],"unable":[10],"to":[11,90,125,138,150,170,178],"reason":[12],"effectively":[13],"in":[14,29],"their":[15],"latent":[16,27,69,92,110,127,141,165],"representations.":[17],"Yet":[18],"little":[19],"is":[20,118],"known":[21],"about":[22],"the":[23,65,108,120,136,148],"limits":[24,34,157],"such":[26],"reasoning":[28],"LLMs.":[30],"We":[31],"test":[32],"these":[33],"by":[35,78],"studying":[36],"whether":[37],"can":[39,114],"discover":[40,86,139],"multi-step":[41],"planning":[42,70,111,166],"strategies":[43,87,161],"without":[44],"supervision":[45,145],"intermediate":[47],"steps":[48,128,167],"and":[49,96,100,147],"execute":[50,151],"them":[51],"latently,":[52],"within":[53],"a":[54,74,133,140],"single":[55],"forward":[56],"pass.":[57],"Using":[58],"graph":[59],"path-finding":[60],"tasks":[61],"that":[62],"precisely":[63],"control":[64],"number":[66],"required":[68],"steps,":[71,93],"we":[72],"uncover":[73],"striking":[75],"limitation":[76],"unresolved":[77],"massive":[79],"scaling:":[80],"tiny":[81],"transformers":[82],"trained":[83],"from":[84],"scratch":[85],"requiring":[88,162],"up":[89,124],"three":[91],"fine-tuned":[94],"GPT-4o":[95],"Qwen3-32B":[97],"reach":[98],"five,":[99,119],"GPT-5.4":[101],"attains":[102],"seven":[103],"under":[104,143],"few-shot":[105],"prompting.":[106],"Although":[107],"maximum":[109],"depth":[112],"learn":[115],"during":[116],"training":[117],"discovered":[121],"strategy":[122,142],"generalizes":[123],"eight":[126],"at":[129],"test-time.":[130],"This":[131],"reveals":[132],"dissociation":[134],"between":[135],"ability":[137,149],"final-answer":[144],"alone":[146],"it":[152],"once":[153],"discovered.":[154],"If":[155],"similar":[156],"hold":[158],"more":[159],"broadly,":[160],"multiple":[163],"coordinated":[164],"may":[168],"need":[169],"be":[171],"explicitly":[172],"taught":[173],"or":[174],"externalized,":[175],"lending":[176],"credence":[177],"CoT":[179],"monitoring.":[180]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-04-10T00:00:00"}