{"id":"W2986265153","doi":"10.1162/coli_a_00367","title":"On the Linguistic Representational Power of Neural Machine Translation Models","year":2020,"lang":"en","type":"preprint","venue":"Computational Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Computer science; Natural language processing; Machine translation; Artificial intelligence; Interpretability; Semantics (computer science); Rule-based machine translation; Word (group theory); Contrast (vision); Linguistics","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004271738,0.000984534,0.0007294641,0.0009573865,0.0005879973,0.002544622,0.001313966,0.001573735,0.002177163],"category_scores_gemma":[0.02894598,0.0006340668,0.0007393414,0.001106649,0.001836443,0.005028845,0.001631675,0.002975306,0.0007315818],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001296785,"about_ca_system_score_gemma":0.0007839593,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005089027,"about_ca_topic_score_gemma":0.004803835,"domain_scores_codex":[0.9983962,0.0009278156,0.00008252268,0.0002550364,0.0002538452,0.00008459042],"domain_scores_gemma":[0.9872896,0.009721479,0.0005635863,0.001447899,0.0008556562,0.0001217276],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002506825,0.00007495966,0.002945811,0.0002172839,0.0001474907,0.0001461853,0.0004006136,0.7520987,0.005330381,0.1041379,0.002280048,0.13197],"study_design_scores_gemma":[0.000007961075,0.00002995369,0.0003561497,0.00003673336,0.00001807575,0.00002778068,0.00002882234,0.9441421,0.00109564,0.05355465,0.0006911599,0.00001095171],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1415518,0.003968054,0.8315929,0.008338322,0.0001519323,0.00005080632,0.0006413536,0.001973169,0.0117316],"genre_scores_gemma":[0.910987,0.001844711,0.08250754,0.0006262787,0.0001819679,0.00008072705,0.0007805653,0.0002231962,0.002767991],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005089027,"threshold_uncertainty_score":0.02259135,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05564198337220849,"score_gpt":0.3233539129337652,"score_spread":0.2677119295615567,"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."}}