{"id":"W4392669708","doi":"10.18653/v1/2023.ijcnlp-main.35","title":"Attacking Open-domain Question Answering by Injecting Misinformation","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"National Science Foundation","keywords":"Misinformation; Chen; Computer science; Domain (mathematical analysis); Question answering; Computational linguistics; Open domain; Natural language processing; Artificial intelligence; Linguistics; Library science; Philosophy; Mathematics; Geology; Computer security","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007616857,0.00006812223,0.00007103987,0.00008701116,0.0001647973,0.0004973824,0.0006447115,0.0000333131,0.000008210439],"category_scores_gemma":[0.00004382435,0.00006768657,0.00001621234,0.0004388595,0.000004777674,0.00183848,0.0005022103,0.00007285642,0.0001206174],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004890894,"about_ca_system_score_gemma":0.00002104514,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001671294,"about_ca_topic_score_gemma":0.0000151959,"domain_scores_codex":[0.9992062,0.00002977951,0.0002011276,0.0001942461,0.0001673704,0.0002012821],"domain_scores_gemma":[0.9995583,0.00004306443,0.00005782389,0.0002714908,0.0000301186,0.00003916013],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000004637254,0.00002689072,0.003648439,0.00007128177,0.00001928417,0.0000126249,0.01085588,0.02421298,0.01900407,0.2439902,0.01510714,0.6830466],"study_design_scores_gemma":[0.0001783376,0.00001433054,0.0004776048,0.00004017639,6.799665e-7,0.000007110481,0.0003052555,0.9881613,0.002825633,0.003847401,0.003999013,0.000143177],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1276455,0.000008586725,0.8606351,0.0006895858,0.0002432916,0.00010788,2.412672e-7,0.0004962436,0.01017358],"genre_scores_gemma":[0.8676641,0.000003922526,0.1311923,0.0003095463,0.00004486643,0.00000948154,0.000007152821,0.00000668122,0.0007619613],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9639483,"threshold_uncertainty_score":0.4796269,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03142707791162644,"score_gpt":0.3068473172510588,"score_spread":0.2754202393394324,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}