{"id":"https://openalex.org/W7220025902","doi":null,"title":"Quantifying Correction Recovery and Robustness Under Prompt Perturbation in Language Model Agents","display_name":"Quantifying Correction Recovery and Robustness Under Prompt Perturbation in Language Model Agents","publication_year":2026,"publication_date":"2026-10-01","ids":{"openalex":"https://openalex.org/W7220025902"},"language":"en","primary_location":{"id":"pmh:oai:ojs2.ojs.zefr.org:article/64","doi":null,"is_oa":true,"landing_page_url":"https://ojs.zefr.org/index.php/intplanet/article/view/64","pdf_url":"https://ojs.zefr.org/index.php/intplanet/article/download/64/60","source":{"id":"https://openalex.org/S7407077709","display_name":"Intelligence Planet Journal of Mathematics and Its Applications","issn_l":"2994-5941","issn":["2994-5941"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2994-5941","raw_type":"info:eu-repo/semantics/article"},"volume":null,"type":"article","is_traditional":true,"indexed_in":[],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://ojs.zefr.org/index.php/intplanet/article/download/64/60","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5037080460","display_name":"Jordan Rowe","orcid":"https://orcid.org/0000-0002-7254-0642","observed_orcids":["https://orcid.org/0000-0002-7254-0642"]},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Rowe, Jordan","raw_affiliation_strings":["The Science Academy STEM Magnet"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Science Academy STEM Magnet","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":["https://openalex.org/A5037080460"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.0,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"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/T12031","display_name":"Speech and dialogue systems","score":0.5928000211715698,"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/T12031","display_name":"Speech and dialogue systems","score":0.5928000211715698,"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":0.15780000388622284,"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/T10906","display_name":"AI-based Problem Solving and Planning","score":0.04919999837875366,"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/language-model-agents","display_name":"language model agents","score":0.9890000224113464},{"id":"https://openalex.org/keywords/tool-calling","display_name":"tool calling","score":0.8759999871253967},{"id":"https://openalex.org/keywords/user-corrections","display_name":"user corrections","score":0.6769999861717224},{"id":"https://openalex.org/keywords/prompt-perturbation","display_name":"prompt perturbation","score":0.9129999876022339},{"id":"https://openalex.org/keywords/robustness","display_name":"robustness","score":0.7509999871253967},{"id":"https://openalex.org/keywords/overcorrection","display_name":"overcorrection","score":0.7609999775886536},{"id":"https://openalex.org/keywords/recovery-cost","display_name":"recovery cost","score":0.765999972820282},{"id":"https://openalex.org/keywords/multi-step-tasks","display_name":"multi-step tasks","score":0.5759999752044678}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6589000225067139},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.5493999719619751},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.5419999957084656},{"id":"https://openalex.org/C48103436","wikidata":"https://www.wikidata.org/wiki/Q599031","display_name":"State (computer science)","level":2,"score":0.3458000123500824},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.3393999934196472},{"id":"https://openalex.org/C177918212","wikidata":"https://www.wikidata.org/wiki/Q803623","display_name":"Perturbation (astronomy)","level":2,"score":0.33219999074935913},{"id":"https://openalex.org/C2776291640","wikidata":"https://www.wikidata.org/wiki/Q2912517","display_name":"Value (mathematics)","level":2,"score":0.3296999931335449},{"id":"https://openalex.org/C79158427","wikidata":"https://www.wikidata.org/wiki/Q485396","display_name":"Analytics","level":2,"score":0.32600000500679016},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.321399986743927},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.319599986076355},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3089999854564667},{"id":"https://openalex.org/C41550386","wikidata":"https://www.wikidata.org/wiki/Q529909","display_name":"Multi-agent system","level":2,"score":0.30630001425743103},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.29589998722076416},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2867000102996826},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.2815000116825104},{"id":"https://openalex.org/C133462117","wikidata":"https://www.wikidata.org/wiki/Q4929239","display_name":"Data collection","level":2,"score":0.2612999975681305},{"id":"https://openalex.org/C113336015","wikidata":"https://www.wikidata.org/wiki/Q574010","display_name":"Complete information","level":2,"score":0.2583000063896179},{"id":"https://openalex.org/C174348530","wikidata":"https://www.wikidata.org/wiki/Q188635","display_name":"Bridging (networking)","level":2,"score":0.25600001215934753},{"id":"https://openalex.org/C2780873155","wikidata":"https://www.wikidata.org/wiki/Q392811","display_name":"Agent-based model","level":2,"score":0.2540000081062317}],"mesh":[],"locations_count":1,"locations":[{"id":"pmh:oai:ojs2.ojs.zefr.org:article/64","doi":null,"is_oa":true,"landing_page_url":"https://ojs.zefr.org/index.php/intplanet/article/view/64","pdf_url":"https://ojs.zefr.org/index.php/intplanet/article/download/64/60","source":{"id":"https://openalex.org/S7407077709","display_name":"Intelligence Planet Journal of Mathematics and Its Applications","issn_l":"2994-5941","issn":["2994-5941"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2994-5941","raw_type":"info:eu-repo/semantics/article"}],"best_oa_location":{"id":"pmh:oai:ojs2.ojs.zefr.org:article/64","doi":null,"is_oa":true,"landing_page_url":"https://ojs.zefr.org/index.php/intplanet/article/view/64","pdf_url":"https://ojs.zefr.org/index.php/intplanet/article/download/64/60","source":{"id":"https://openalex.org/S7407077709","display_name":"Intelligence Planet Journal of Mathematics and Its Applications","issn_l":"2994-5941","issn":["2994-5941"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2994-5941","raw_type":"info:eu-repo/semantics/article"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7220025902.pdf","grobid_xml":"https://content.openalex.org/works/W7220025902.grobid-xml"},"referenced_works_count":3,"referenced_works":["https://openalex.org/W4304195432","https://openalex.org/W4399836817","https://openalex.org/W4416055245"],"related_works":["https://openalex.org/W7220930860","https://openalex.org/W7127200483","https://openalex.org/W4415338087","https://openalex.org/W7204258167","https://openalex.org/W7130720879","https://openalex.org/W7171639259","https://openalex.org/W7154139717","https://openalex.org/W7210269164","https://openalex.org/W7214735499","https://openalex.org/W7213355416"],"abstract_inverted_index":{"Language":[0],"model":[1],"agents":[2],"that":[3,12,137],"call":[4],"external":[5],"tools":[6],"now":[7],"carry":[8],"out":[9],"multi-step":[10],"tasks":[11],"change":[13],"real":[14],"state,":[15],"booking":[16],"rooms,":[17],"scheduling":[18],"meetings,":[19],"placing":[20],"orders.":[21],"In":[22,60],"practice,":[23],"people":[24],"revise":[25],"their":[26],"intent":[27],"partway":[28],"through,":[29],"correcting":[30],"a":[31,46,51,65,72,133,144,239,254],"value":[32,167],"they":[33],"gave":[34],"earlier.":[35],"Whether":[36],"the":[37,56,80,85,99,105,107,117,141,158,165,179,217],"agent":[38,57,81,108,142,203,214],"then":[39],"recovers":[40],"is":[41,77,89,109,114,168,198,236],"rarely":[42],"measured,":[43],"even":[44],"if":[45],"failed":[47],"recovery":[48,94,125,127,258],"can":[49,95],"leave":[50],"wrong":[52,166],"action":[53],"committed":[54,100,170],"while":[55],"reports":[58],"success.":[59],"this":[61],"work,":[62],"we":[63],"propose":[64],"framework":[66,121,197],"to":[67,82,189,223],"measure":[68],"what":[69,84,115],"happens":[70],"after":[71],"user":[73,210],"correction.":[74],"Every":[75],"task":[76],"scripted":[78,232],"for":[79],"know":[83],"correct":[86],"final":[87],"state":[88,102],"before":[90,246],"it":[91],"runs,":[92],"so":[93,216],"be":[96],"measured":[97],"against":[98],"tool":[101,262],"rather":[103,171],"than":[104,172],"text":[106],"writing.":[110],"That":[111],"single":[112],"choice":[113],"makes":[116],"measurement":[118],"possible.":[119],"The":[120,196,234],"defines":[122],"four":[123],"quantities:":[124],"success,":[126],"cost,":[128],"an":[129,202],"overcorrection":[130],"rate,":[131],"and":[132,243,249],"silent":[134],"inconsistency":[135],"rate":[136],"counts":[138],"runs":[139],"where":[140],"says":[143],"fix":[145],"never":[146],"happened.":[147],"Four":[148],"hypotheses":[149],"are":[150,183,193,221,251],"fixed":[151,245],"in":[152,212,261],"advance.":[153],"Recovery":[154],"cost":[155],"rises":[156,163],"when":[157,164],"correction":[159,260],"arrives":[160],"later.":[161],"Cost":[162],"already":[169],"just":[173],"examined.":[174],"A":[175],"significant":[176],"percentage":[177],"of":[178,259],"seemingly":[180],"successful":[181],"corrections":[182,211],"quietly":[184],"wrong.":[185],"Agents":[186],"more":[187],"sensitive":[188],"small":[190],"prompt":[191],"changes":[192],"worse":[194],"off.":[195],"built":[199],"on":[200,225,231,257],"Voker,":[201],"analytics":[204],"platform":[205],"whose":[206],"SDK":[207],"automatically":[208],"detects":[209],"production":[213],"conversations,":[215],"measurements":[218],"defined":[219],"here":[220],"meant":[222],"run":[224],"live":[226],"traffic":[227],"as":[228,230,238,253],"well":[229],"runs.":[233],"study":[235],"written":[237],"preregistration":[240],"with":[241],"method":[242],"predictions":[244],"data":[247],"collection,":[248],"results":[250],"available":[252],"working":[255],"harness":[256],"calling":[263],"agents.":[264]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-10-05T00:00:00"}