{"id":"https://openalex.org/W7202065664","doi":null,"title":"Adversarial Attacks on Code Large Language Models Using Reinforcement Learning","display_name":"Adversarial Attacks on Code Large Language Models Using Reinforcement Learning","publication_year":2026,"publication_date":"2026-07-17","ids":{"openalex":"https://openalex.org/W7202065664"},"language":"en","primary_location":{"id":"pmh:oai:repositorio-aberto.up.pt:10216/176075","doi":null,"is_oa":true,"landing_page_url":"https://hdl.handle.net/10216/176075","pdf_url":"https://repositorio-aberto.up.pt/bitstream/10216/176075/2/787332.pdf","source":{"id":"https://openalex.org/S4306402433","display_name":"Portuguese National Funding Agency for Science, Research and Technology (RCAAP Project by FCT)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"master thesis"},"volume":null,"type":"dissertation","is_traditional":true,"indexed_in":[],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://repositorio-aberto.up.pt/bitstream/10216/176075/2/787332.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5146203277","display_name":"Pedro Jorge Mendes Jesus Landolt","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Pedro Jorge Mendes Jesus Landolt","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":["https://openalex.org/A5146203277"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9375,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9375,"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/T10743","display_name":"Software Testing and Debugging Techniques","score":0.03579999879002571,"subfield":{"id":"https://openalex.org/subfields/1712","display_name":"Software"},"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/T10260","display_name":"Software Engineering Research","score":0.016100000590085983,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/adversarial-attacks","display_name":"adversarial attacks","score":0.9380000233650208},{"id":"https://openalex.org/keywords/large-language-models-for-code","display_name":"code large language models","score":0.875},{"id":"https://openalex.org/keywords/code-review","display_name":"code review","score":0.6859999895095825},{"id":"https://openalex.org/keywords/llm-as-a-judge","display_name":"LLM-as-a-judge","score":0.5429999828338623},{"id":"https://openalex.org/keywords/multi-armed-bandits","display_name":"multi-armed bandits","score":0.7940000295639038},{"id":"https://openalex.org/keywords/black-box-attacks","display_name":"black-box attacks","score":0.6320000290870667},{"id":"https://openalex.org/keywords/prompt-injection","display_name":"prompt injection","score":0.6460000276565552},{"id":"https://openalex.org/keywords/unit-testing","display_name":"unit testing","score":0.48500001430511475},{"id":"https://openalex.org/keywords/jester","display_name":"jester","score":0.5979999899864197}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5364000201225281},{"id":"https://openalex.org/C15708023","wikidata":"https://www.wikidata.org/wiki/Q80083","display_name":"Humanities","level":1,"score":0.4625999927520752},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.4603999853134155},{"id":"https://openalex.org/C519991488","wikidata":"https://www.wikidata.org/wiki/Q28865","display_name":"Python (programming language)","level":2,"score":0.4165000021457672},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.39649999141693115},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.3788999915122986},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.3564999997615814},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.3140000104904175},{"id":"https://openalex.org/C205649164","wikidata":"https://www.wikidata.org/wiki/Q1071","display_name":"Geography","level":0,"score":0.2621000111103058},{"id":"https://openalex.org/C48372109","wikidata":"https://www.wikidata.org/wiki/Q3913","display_name":"Binary number","level":2,"score":0.2619999945163727},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.25029999017715454}],"mesh":[],"locations_count":1,"locations":[{"id":"pmh:oai:repositorio-aberto.up.pt:10216/176075","doi":null,"is_oa":true,"landing_page_url":"https://hdl.handle.net/10216/176075","pdf_url":"https://repositorio-aberto.up.pt/bitstream/10216/176075/2/787332.pdf","source":{"id":"https://openalex.org/S4306402433","display_name":"Portuguese National Funding Agency for Science, Research and Technology (RCAAP Project by FCT)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"master thesis"}],"best_oa_location":{"id":"pmh:oai:repositorio-aberto.up.pt:10216/176075","doi":null,"is_oa":true,"landing_page_url":"https://hdl.handle.net/10216/176075","pdf_url":"https://repositorio-aberto.up.pt/bitstream/10216/176075/2/787332.pdf","source":{"id":"https://openalex.org/S4306402433","display_name":"Portuguese National Funding Agency for Science, Research and Technology (RCAAP Project by FCT)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"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":"master thesis"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7202065664.pdf","grobid_xml":"https://content.openalex.org/works/W7202065664.grobid-xml"},"referenced_works_count":32,"referenced_works":["https://openalex.org/W1501823362","https://openalex.org/W2039522160","https://openalex.org/W3030163527","https://openalex.org/W3034549508","https://openalex.org/W3177813494","https://openalex.org/W3195140297","https://openalex.org/W4206547457","https://openalex.org/W4226278401","https://openalex.org/W4304195432","https://openalex.org/W4320854981","https://openalex.org/W4386185625","https://openalex.org/W4386436506","https://openalex.org/W4386839891","https://openalex.org/W4391272793","https://openalex.org/W4401306886","https://openalex.org/W4403796339","https://openalex.org/W4403853639","https://openalex.org/W4406093459","https://openalex.org/W4406491467","https://openalex.org/W4406959316","https://openalex.org/W4407601344","https://openalex.org/W4407783204","https://openalex.org/W4409360937","https://openalex.org/W4414972656","https://openalex.org/W4415160894","https://openalex.org/W4416251571","https://openalex.org/W4416411532","https://openalex.org/W6929374590","https://openalex.org/W6967084484","https://openalex.org/W7139945551","https://openalex.org/W7164823666","https://openalex.org/W7166827396"],"related_works":["https://openalex.org/W4412509569","https://openalex.org/W2980634774","https://openalex.org/W4417295848","https://openalex.org/W3090259656","https://openalex.org/W7166559797","https://openalex.org/W3034846000","https://openalex.org/W7166822288","https://openalex.org/W4382317573","https://openalex.org/W7130732740","https://openalex.org/W4416380752"],"abstract_inverted_index":{"Code":[0],"review":[1],"increasingly":[2],"relies":[3],"on":[4,90,188,192],"LLMs":[5],"to":[6,53,57,98,249],"evaluate":[7,186],"submitted":[8],"code.":[9],"A":[10,83],"function":[11],"that":[12,33,59,92,110,176,180],"fails":[13],"every":[14],"deterministic":[15,164],"unit":[16],"test":[17,43],"can":[18],"still":[19],"pass":[20],"an":[21,25],"LLM":[22],"judge":[23,97,120,202,215],"when":[24],"attacker":[26],"presents":[27],"it":[28],"with":[29,162],"a":[30,79,107,130,133,138,156,173],"crafted":[31],"framing":[32,51],"exploits":[34],"reasoning":[35,135],"weaknesses":[36],"in":[37,257],"the":[38,87,96,193,213,218,227,237,241,246,253],"model,":[39,216],"such":[40],"as":[41,78],"fabricated":[42],"output,":[44],"contextual":[45],"pressure,":[46],"or":[47],"apparent":[48],"authority.":[49],"Which":[50],"strategy":[52],"apply,":[54],"and":[55,119,137,149,153,169,191,240,252],"how":[56],"adapt":[58],"choice":[60],"across":[61,142,199],"repeated":[62],"attempts,":[63],"remain":[64],"open":[65],"questions.":[66],"JESTER":[67,105,187,242],"(Judge":[68],"Evaluation":[69],"via":[70],"Selective":[71],"Tactic-based":[72],"Exploit":[73],"Reinforcement)":[74],"treats":[75],"tactic":[76,254],"selection":[77,128],"sequential":[80],"decision":[81],"problem.":[82],"multi-armed":[84],"bandit":[85,139],"concentrates":[86],"evaluation":[88,121],"budget":[89],"framings":[91],"have":[93],"historically":[94],"caused":[95],"fail,":[99],"operating":[100],"under":[101],"fully":[102],"black-box":[103],"constraints.":[104],"implements":[106],"four-step":[108],"pipeline":[109,175],"separates":[111],"bug":[112],"generation,":[113],"ground":[114],"truth":[115],"validation,":[116],"adversarial":[117],"framing,":[118],"into":[122],"independent":[123],"stages.":[124],"I":[125,151,185,235],"compare":[126],"three":[127],"policies:":[129],"random":[131,219],"baseline,":[132],"ReAct-based":[134],"selector,":[136],"selector":[140],"instantiated":[141],"four":[143,200],"algorithms":[144],"(UCB1,":[145],"Thompson":[146],"Sampling,":[147],"KL-UCB,":[148],"EXP3).":[150],"construct":[152],"publish":[154,236],"adversarial-code-buggy,":[155],"dataset":[157,190,239],"of":[158,221],"982":[159],"Python":[160],"functions":[161],"confirmed":[163],"bugs":[165],"drawn":[166],"from":[167],"MBPP":[168],"HumanEval,":[170],"filtered":[171],"by":[172],"two-judge":[174],"retains":[177],"only":[178],"records":[179],"are":[181],"genuine":[182],"attack":[183,207,230],"targets.":[184],"this":[189],"CuBERT":[194],"Wrong":[195],"Binary":[196],"Operator":[197],"corpus":[198],"open-weight":[201],"models.":[203],"KL-UCB":[204],"achieves":[205,226],"67.6%":[206],"success":[208,231],"at":[209,233],"first":[210],"attempt":[211],"against":[212],"primary":[214],"doubling":[217],"baseline":[220],"35.1%,":[222],"while":[223],"Prompt":[224],"Injection":[225],"highest":[228],"single-tactic":[229],"rate":[232],"62.8%.":[234],"adversarial-code-buggy":[238],"codebase":[243],"openly,":[244],"allowing":[245],"experimental":[247],"conditions":[248],"be":[250],"reproduced":[251],"registry":[255],"extended":[256],"future":[258],"work.":[259]},"counts_by_year":[],"updated_date":"2026-11-08T12:12:28.789561","created_date":"2026-08-10T00:00:00"}