{"id":"https://openalex.org/W7204192724","doi":"https://doi.org/10.48550/arxiv.2608.21664","title":"Measuring Activation Control in Large Language Models","display_name":"Measuring Activation Control in Large Language Models","publication_year":2026,"publication_date":"2026-08-21","ids":{"openalex":"https://openalex.org/W7204192724","doi":"https://doi.org/10.48550/arxiv.2608.21664"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.21664","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.21664","pdf_url":"https://arxiv.org/pdf/2608.21664","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2608.21664","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5077146802","display_name":"Marek Mateusz Kowalski","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kowalski, Marek Mateusz","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5141582893","display_name":"Joshua Fonseca Rivera","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rivera, Joshua Fonseca","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5039029924","display_name":"Uzay Macar","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Macar, Uzay","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5119181839","display_name":"David Demitri Africa","orcid":"https://orcid.org/0009-0005-3543-8891","observed_orcids":["https://orcid.org/0009-0005-3543-8891"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Africa, David Demitri","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":true,"cited_by_count":2,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.8389000296592712,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.8389000296592712,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.13379999995231628,"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/T10883","display_name":"Ethics and Social Impacts of AI","score":0.012900000438094139,"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/large-language-models","display_name":"large language models","score":0.9129999876022339},{"id":"https://openalex.org/keywords/residual-stream","display_name":"residual stream","score":0.8849999904632568},{"id":"https://openalex.org/keywords/linear-array-transducer","display_name":"linear array transducer","score":0.5529999732971191},{"id":"https://openalex.org/keywords/model-introspection","display_name":"model introspection","score":0.6169999837875366},{"id":"https://openalex.org/keywords/deception","display_name":"deception","score":0.45100000500679016}],"concepts":[{"id":"https://openalex.org/C48209547","wikidata":"https://www.wikidata.org/wiki/Q1331104","display_name":"Controllability","level":2,"score":0.6837999820709229},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6658999919891357},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.5044000148773193},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.48750001192092896},{"id":"https://openalex.org/C112313634","wikidata":"https://www.wikidata.org/wiki/Q7886648","display_name":"Complement (music)","level":5,"score":0.4779999852180481},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.450300008058548},{"id":"https://openalex.org/C155512373","wikidata":"https://www.wikidata.org/wiki/Q287450","display_name":"Residual","level":2,"score":0.43560001254081726},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.4341000020503998},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.38760000467300415},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.3601999878883362},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.3544999957084656},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3476000130176544},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.34040001034736633},{"id":"https://openalex.org/C12713177","wikidata":"https://www.wikidata.org/wiki/Q1900281","display_name":"Perspective (graphical)","level":2,"score":0.3192000091075897},{"id":"https://openalex.org/C2779628075","wikidata":"https://www.wikidata.org/wiki/Q1253258","display_name":"Downgrade","level":2,"score":0.30149999260902405},{"id":"https://openalex.org/C75291252","wikidata":"https://www.wikidata.org/wiki/Q1315756","display_name":"TRACE (psycholinguistics)","level":2,"score":0.28459998965263367},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.26930001378059387},{"id":"https://openalex.org/C31531917","wikidata":"https://www.wikidata.org/wiki/Q915157","display_name":"Robust control","level":3,"score":0.2597000002861023},{"id":"https://openalex.org/C94966114","wikidata":"https://www.wikidata.org/wiki/Q29256","display_name":"Black box","level":2,"score":0.25619998574256897},{"id":"https://openalex.org/C47446073","wikidata":"https://www.wikidata.org/wiki/Q5165890","display_name":"Control theory (sociology)","level":3,"score":0.25369998812675476},{"id":"https://openalex.org/C166998942","wikidata":"https://www.wikidata.org/wiki/Q16969238","display_name":"Interactor","level":2,"score":0.2535000145435333},{"id":"https://openalex.org/C177774035","wikidata":"https://www.wikidata.org/wiki/Q1246948","display_name":"Granularity","level":2,"score":0.2533000111579895}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.21664","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.21664","pdf_url":"https://arxiv.org/pdf/2608.21664","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2608.21664","doi":"https://doi.org/10.48550/arxiv.2608.21664","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.21664","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:2608.21664","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.21664","pdf_url":"https://arxiv.org/pdf/2608.21664","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7204192724.pdf","grobid_xml":"https://content.openalex.org/works/W7204192724.grobid-xml"},"referenced_works_count":16,"referenced_works":["https://openalex.org/W2529194139","https://openalex.org/W2948625193","https://openalex.org/W4387225729","https://openalex.org/W4405355489","https://openalex.org/W4415971848","https://openalex.org/W4416789529","https://openalex.org/W4417447071","https://openalex.org/W6910518143","https://openalex.org/W7116043699","https://openalex.org/W7118379034","https://openalex.org/W7119898775","https://openalex.org/W7124717434","https://openalex.org/W7131371221","https://openalex.org/W7134055960","https://openalex.org/W7140204093","https://openalex.org/W7166813420"],"related_works":["https://openalex.org/W6910518143","https://openalex.org/W7222065977","https://openalex.org/W7214888684","https://openalex.org/W7171637477","https://openalex.org/W7117292313","https://openalex.org/W4417356736","https://openalex.org/W7106653175","https://openalex.org/W4417168559","https://openalex.org/W7116400519","https://openalex.org/W4416548317"],"abstract_inverted_index":{"Safe":[0],"deployment":[1],"of":[2,87,95,109],"increasingly":[3],"capable":[4],"models":[5,23,30,61],"will":[6],"likely":[7],"come":[8],"to":[9,17,55,59],"rely":[10],"on":[11],"latent-space":[12],"monitoring":[13,114,145],"as":[14,146],"a":[15,142],"complement":[16],"behavioral":[18],"evaluations,":[19],"especially":[20],"when":[21],"evaluation-aware":[22],"exhibit":[24],"scheming":[25],"or":[26],"deception.":[27],"However,":[28],"if":[29],"can":[31,62,81,111],"also":[32],"control":[33,82,110,134],"their":[34,64,88],"own":[35],"activations,":[36],"deception":[37],"could":[38,140],"extend":[39],"into":[40],"the":[41,51,57,83,125,136],"latent":[42],"space":[43,138],"itself.":[44],"With":[45],"this":[46,107],"in":[47,161],"mind,":[48],"we":[49,76,151],"introduce":[50],"Activation":[52],"Controllability":[53],"Benchmark":[54],"quantify":[56],"extent":[58],"which":[60],"modulate":[63],"residual":[65,89],"stream":[66,90],"via":[67],"natural-language":[68],"instruction.":[69],"Across":[70],"model":[71],"families":[72],"and":[73,85,124,156],"capability":[74],"levels,":[75],"find":[77],"that":[78,133,153],"most":[79],"LLMs":[80],"direction":[84],"magnitude":[86],"activations":[91],"with":[92],"some":[93],"degree":[94],"temporal":[96],"resolution,":[97],"though":[98],"performance":[99],"varies":[100],"considerably":[101],"across":[102],"models.":[103,163],"In":[104],"simple":[105],"tasks,":[106],"level":[108],"evade":[112],"activation-based":[113],"methods":[115],"(including":[116],"linear":[117],"probes,":[118],"natural":[119],"language":[120],"autoencoders,":[121],"activation":[122,137,159],"oracles,":[123],"Jacobian":[126],"lens),":[127],"albeit":[128],"imperfectly.":[129],"These":[130],"results":[131],"suggest":[132],"over":[135],"itself":[139],"become":[141],"confound":[143],"for":[144],"introspective":[147],"capabilities":[148],"increase;":[149],"therefore,":[150],"recommend":[152],"frontier":[154],"labs":[155],"evaluators":[157],"track":[158],"controllability":[160],"future":[162]},"counts_by_year":[{"year":2026,"cited_by_count":2}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-08-26T00:00:00"}