{"id":"https://openalex.org/W4415091089","doi":"https://doi.org/10.48550/arxiv.2509.10625","title":"No Answer Needed: Predicting LLM Answer Accuracy from Question-Only Linear Probes","display_name":"No Answer Needed: Predicting LLM Answer Accuracy from Question-Only Linear Probes","publication_year":2025,"publication_date":"2025-09-12","ids":{"openalex":"https://openalex.org/W4415091089","doi":"https://doi.org/10.48550/arxiv.2509.10625"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2509.10625","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2509.10625","pdf_url":"https://arxiv.org/pdf/2509.10625","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-sa","license_id":"https://openalex.org/licenses/cc-by-sa","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/2509.10625","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5119959931","display_name":"Iv\u00e1n Vicente Moreno Cencerrado","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cencerrado, Iv\u00e1n Vicente Moreno","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5119959932","display_name":"Arnau Padr\u00e9s Masdemont","orcid":"https://orcid.org/0009-0004-9907-3548","observed_orcids":["https://orcid.org/0009-0004-9907-3548"]},"institutions":[{"id":"https://openalex.org/I241749","display_name":"University of Cambridge","ror":"https://ror.org/013meh722","country_code":"GB","type":"education","lineage":["https://openalex.org/I241749"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Masdemont, Arnau Padr\u00e9s","raw_affiliation_strings":["University of Cambridge , MARS"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Cambridge , MARS","institution_ids":["https://openalex.org/I241749"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5119959933","display_name":"Anton Gonzalvez Hawthorne","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I2802026069","display_name":"Valencian International University","ror":"https://ror.org/00gjj5n39","country_code":"ES","type":"education","lineage":["https://openalex.org/I2802026069"]}],"countries":["ES"],"is_corresponding":false,"raw_author_name":"Hawthorne, Anton Gonzalvez","raw_affiliation_strings":["Universidad Internacional de Valencia , MARS"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Universidad Internacional de Valencia , MARS","institution_ids":["https://openalex.org/I2802026069"]}]},{"author_position":"middle","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":[{"id":"https://openalex.org/I241749","display_name":"University of Cambridge","ror":"https://ror.org/013meh722","country_code":"GB","type":"education","lineage":["https://openalex.org/I241749"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Africa, David Demitri","raw_affiliation_strings":["University of Cambridge , MARS"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Cambridge , MARS","institution_ids":["https://openalex.org/I241749"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5018689327","display_name":"Lorenzo Pacchiardi","orcid":"https://orcid.org/0000-0003-4760-7638","observed_orcids":["https://orcid.org/0000-0003-4760-7638"]},"institutions":[{"id":"https://openalex.org/I241749","display_name":"University of Cambridge","ror":"https://ror.org/013meh722","country_code":"GB","type":"education","lineage":["https://openalex.org/I241749"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Pacchiardi, Lorenzo","raw_affiliation_strings":["University of Cambridge"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Cambridge","institution_ids":["https://openalex.org/I241749"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":39,"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":0.49779999256134033,"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":0.49779999256134033,"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":0.44290000200271606,"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":0.02590000070631504,"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":0.9150000214576721},{"id":"https://openalex.org/keywords/linear-array-transducer","display_name":"linear array transducer","score":0.6990000009536743},{"id":"https://openalex.org/keywords/trivia-questions","display_name":"trivia questions","score":0.5740000009536743},{"id":"https://openalex.org/keywords/out-of-distribution-generalization","display_name":"out-of-distribution generalization","score":0.8899999856948853},{"id":"https://openalex.org/keywords/mathematical-reasoning","display_name":"mathematical reasoning","score":0.9190000295639038},{"id":"https://openalex.org/keywords/confidence-estimation","display_name":"confidence estimation","score":0.5070000290870667},{"id":"https://openalex.org/keywords/spurious-features","display_name":"spurious features","score":0.47699999809265137}],"concepts":[{"id":"https://openalex.org/C97256817","wikidata":"https://www.wikidata.org/wiki/Q1462316","display_name":"Spurious relationship","level":2,"score":0.7960000038146973},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.638700008392334},{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.6133000254631042},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5810999870300293},{"id":"https://openalex.org/C2778136018","wikidata":"https://www.wikidata.org/wiki/Q10350689","display_name":"Predictive power","level":2,"score":0.54830002784729},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.49309998750686646},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.4196999967098236},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.41260001063346863},{"id":"https://openalex.org/C163258240","wikidata":"https://www.wikidata.org/wiki/Q25342","display_name":"Power (physics)","level":2,"score":0.39149999618530273},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.3361999988555908},{"id":"https://openalex.org/C163175372","wikidata":"https://www.wikidata.org/wiki/Q3339222","display_name":"Linear model","level":2,"score":0.3325999975204468},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.2750000059604645},{"id":"https://openalex.org/C149441793","wikidata":"https://www.wikidata.org/wiki/Q200726","display_name":"Probability distribution","level":2,"score":0.2581000030040741},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.2563000023365021}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2509.10625","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2509.10625","pdf_url":"https://arxiv.org/pdf/2509.10625","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-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2509.10625","doi":"https://doi.org/10.48550/arxiv.2509.10625","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2509.10625","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2509.10625","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2509.10625","pdf_url":"https://arxiv.org/pdf/2509.10625","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-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4415091089.pdf","grobid_xml":"https://content.openalex.org/works/W4415091089.grobid-xml"},"referenced_works_count":25,"referenced_works":["https://openalex.org/W2612431505","https://openalex.org/W3035441651","https://openalex.org/W4285429195","https://openalex.org/W4310926773","https://openalex.org/W4321594373","https://openalex.org/W4322718191","https://openalex.org/W4388514540","https://openalex.org/W4389518688","https://openalex.org/W4402670143","https://openalex.org/W4403556531","https://openalex.org/W4404752296","https://openalex.org/W4404782975","https://openalex.org/W4405431306","https://openalex.org/W4406092806","https://openalex.org/W4407129404","https://openalex.org/W4407231841","https://openalex.org/W4411121371","https://openalex.org/W4412888163","https://openalex.org/W4412888889","https://openalex.org/W4414934008","https://openalex.org/W4416100856","https://openalex.org/W4416150232","https://openalex.org/W6929465362","https://openalex.org/W7126220794","https://openalex.org/W7130610581"],"related_works":["https://openalex.org/W7165115828","https://openalex.org/W7140074841","https://openalex.org/W7053377175","https://openalex.org/W7154479124","https://openalex.org/W7167933819","https://openalex.org/W4412888889","https://openalex.org/W4415337611","https://openalex.org/W4416373935","https://openalex.org/W7215133724","https://openalex.org/W4416339328"],"abstract_inverted_index":{"Do":[0],"large":[1],"language":[2],"models":[3,108],"(LLMs)":[4],"anticipate":[5],"when":[6],"they":[7],"will":[8,39],"answer":[9,38],"correctly?":[10],"To":[11],"study":[12],"this,":[13],"we":[14],"extract":[15],"activations":[16],"after":[17],"a":[18,76],"question":[19],"is":[20],"read":[21],"but":[22],"before":[23],"any":[24],"tokens":[25],"are":[26],"generated,":[27],"and":[28,69,83,87,135,141],"train":[29],"linear":[30],"probes":[31,140],"to":[32,50,149],"predict":[33,65],"whether":[34],"the":[35,118,123],"model's":[36],"forthcoming":[37],"be":[40],"correct.":[41],"Across":[42],"three":[43],"open-source":[44],"model":[45],"families":[46],"ranging":[47],"from":[48],"7":[49],"70":[51],"billion":[52],"parameters,":[53],"projections":[54],"on":[55,61,70,101,133],"this":[56],"\"in-advance":[57],"correctness":[58],"direction\"":[59],"trained":[60],"generic":[62],"trivia":[63],"questions":[64,102],"success":[66],"in":[67,94],"distribution":[68],"diverse":[71],"out-of-distribution":[72],"knowledge":[73],"datasets,":[74],"indicating":[75,121],"deeper":[77],"signal":[78],"than":[79],"dataset-specific":[80],"spurious":[81],"features,":[82],"outperforming":[84],"black-box":[85],"baselines":[86],"verbalised":[88],"predicted":[89],"confidence.":[90,128],"Predictive":[91],"power":[92],"saturates":[93],"intermediate":[95],"layers":[96],"and,":[97],"notably,":[98],"generalisation":[99],"falters":[100],"requiring":[103],"mathematical":[104],"reasoning.":[105],"Moreover,":[106],"for":[107],"responding":[109],"\"I":[110],"don't":[111],"know\",":[112],"doing":[113],"so":[114],"strongly":[115],"correlates":[116],"with":[117,139],"probe":[119],"score,":[120],"that":[122],"same":[124],"direction":[125],"also":[126],"captures":[127],"By":[129],"complementing":[130],"previous":[131],"results":[132],"truthfulness":[134],"other":[136],"behaviours":[137],"obtained":[138],"sparse":[142],"auto-encoders,":[143],"our":[144],"work":[145],"contributes":[146],"essential":[147],"findings":[148],"elucidate":[150],"LLM":[151],"internals.":[152]},"counts_by_year":[{"year":2026,"cited_by_count":36},{"year":2025,"cited_by_count":3}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2025-10-13T00:00:00"}