{"id":"W4391833165","doi":"10.1145/3640543.3645200","title":"Why and When LLM-Based Assistants Can Go Wrong: Investigating the Effectiveness of Prompt-Based Interactions for Software Help-Seeking","year":2024,"lang":"en","type":"preprint","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":61,"is_retracted":false,"has_abstract":true,"ca_institutions":"Simon Fraser University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Context (archaeology); Task (project management); Domain (mathematical analysis); Software; Baseline (sea); Relevance (law); Perception; Software engineering; Psychology; Engineering; Programming language","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01100309,0.0008695642,0.0006817722,0.0008852049,0.0005802698,0.001880136,0.001224725,0.0015633,0.001759996],"category_scores_gemma":[0.1176521,0.0005500543,0.0004181502,0.0003387954,0.0007841079,0.00283979,0.001829763,0.001010998,0.0009794631],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005224212,"about_ca_system_score_gemma":0.0009874047,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008993261,"about_ca_topic_score_gemma":0.001307884,"domain_scores_codex":[0.9891554,0.007523718,0.0006376667,0.001134899,0.001167027,0.0003813904],"domain_scores_gemma":[0.8689678,0.1107764,0.008460506,0.005400531,0.004324922,0.00206983],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.01091834,0.007074443,0.1289231,0.006418298,0.0003659888,0.002127312,0.121962,0.005658621,0.1393054,0.001841506,0.00508199,0.5703231],"study_design_scores_gemma":[0.003598591,0.06318893,0.4266435,0.002949591,0.001860228,0.006611443,0.09075286,0.1706809,0.164195,0.01167038,0.05644555,0.001402965],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9820772,0.0002326897,0.01479201,0.0001501458,0.00002530678,0.0002940733,0.0000755993,0.001140591,0.001212247],"genre_scores_gemma":[0.9558839,0.000119825,0.04206684,0.000238367,0.00002381412,0.0003980868,0.0001329651,0.0001489796,0.0009872826],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01100309,"threshold_uncertainty_score":0.05819058,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04492530613574788,"score_gpt":0.2926778040444654,"score_spread":0.2477524979087175,"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."}}