{"id":"W4406499798","doi":"10.1109/cascon62161.2024.10838185","title":"Sentiment Analysis with LLMs: Evaluating QLoRA Fine-Tuning, Instruction Strategies, and Prompt Sensitivity","year":2024,"lang":"en","type":"article","venue":"","topic":"Sentiment Analysis and Opinion Mining","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"Toronto Metropolitan University","funders":"","keywords":"Sensitivity (control systems); Computer science; Sentiment analysis; Artificial intelligence; Engineering","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":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0007646747,0.0001590306,0.00022321,0.0004032093,0.0001598849,0.001500652,0.0001082943,0.00003597304,0.00008425382],"category_scores_gemma":[0.000007234017,0.0001174348,0.00009142554,0.001700281,0.00003802004,0.001182695,0.0001321055,0.0001092195,0.00001314981],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003630367,"about_ca_system_score_gemma":0.00009719696,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001489675,"about_ca_topic_score_gemma":0.0001773186,"domain_scores_codex":[0.9984526,0.0001120839,0.0002264448,0.0005772554,0.0004267006,0.0002049178],"domain_scores_gemma":[0.9993988,0.00007465306,0.00007048618,0.0003018982,0.0000808402,0.00007329856],"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":[0.00004469225,0.0002375492,0.05024355,0.00027958,0.01029422,0.0004879137,0.009741994,0.04978089,0.01807671,0.2254887,0.00102812,0.6342961],"study_design_scores_gemma":[0.0001649651,0.00008659829,0.004250796,0.00004277387,0.0003587044,0.00002912248,0.0006303561,0.9929435,0.0008975018,0.0002263636,0.0001872884,0.0001820211],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.492185,0.0002002297,0.5057405,0.0003386956,0.0001030426,0.00009030852,6.142232e-7,0.0001920408,0.001149612],"genre_scores_gemma":[0.9501826,0.000007054013,0.04920036,0.00003079943,0.00005480317,0.000006299279,0.00001127092,0.000006631254,0.000500203],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9431626,"threshold_uncertainty_score":0.9995359,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0271359459686862,"score_gpt":0.2991807555376677,"score_spread":0.2720448095689815,"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."}}