{ "handler": { "task": "handler", "best_epoch": 5, "training_seconds": 304.73208945803344, "split_sizes": { "train": 2400, "validation": 300, "test": 300 }, "validation": { "loss": 0.5876449354489645, "accuracy": 0.8466666666666667, "macro_f1": 0.8471922277069336 }, "test": { "loss": 0.6294241340955099, "accuracy": 0.8166666666666667, "macro_f1": 0.816800011054457 }, "test_confusions": [ { "gold": "human", "predicted": "llm", "count": 18 }, { "gold": "llm", "predicted": "auto", "count": 15 }, { "gold": "llm", "predicted": "human", "count": 10 }, { "gold": "auto", "predicted": "llm", "count": 5 }, { "gold": "auto", "predicted": "human", "count": 5 }, { "gold": "human", "predicted": "auto", "count": 2 } ], "history": [ { "epoch": 1, "train_loss": 0.9450590944290161, "validation_loss": 0.640375197728475, "validation_accuracy": 0.7833333333333333, "validation_macro_f1": 0.7873895322407908 }, { "epoch": 2, "train_loss": 0.3881728690862656, "validation_loss": 0.5689977049827576, "validation_accuracy": 0.8366666666666667, "validation_macro_f1": 0.8295788111791534 }, { "epoch": 3, "train_loss": 0.18638452569643657, "validation_loss": 0.6046478613217672, "validation_accuracy": 0.8033333333333333, "validation_macro_f1": 0.8073069144063059 }, { "epoch": 4, "train_loss": 0.17458565334479015, "validation_loss": 0.5988235028584799, "validation_accuracy": 0.8333333333333334, "validation_macro_f1": 0.8331806628375255 }, { "epoch": 5, "train_loss": 0.17275874733924865, "validation_loss": 0.5876449354489645, "validation_accuracy": 0.8466666666666667, "validation_macro_f1": 0.8471922277069336 }, { "epoch": 6, "train_loss": 0.17210762143135072, "validation_loss": 0.5890995128949483, "validation_accuracy": 0.84, "validation_macro_f1": 0.839868859643578 }, { "epoch": 7, "train_loss": 0.17215188225110373, "validation_loss": 0.5889288687705994, "validation_accuracy": 0.84, "validation_macro_f1": 0.8401046298542054 }, { "epoch": 8, "train_loss": 0.1719669226805369, "validation_loss": 0.5910564653078715, "validation_accuracy": 0.8366666666666667, "validation_macro_f1": 0.8366387893269615 } ] }, "human": { "task": "human", "best_epoch": 6, "training_seconds": 111.95542904100148, "split_sizes": { "train": 880, "validation": 110, "test": 110 }, "validation": { "loss": 1.267178336056796, "accuracy": 0.8363636363636363, "macro_f1": 0.7900596877869606 }, "test": { "loss": 1.4121767195788297, "accuracy": 0.7454545454545455, "macro_f1": 0.6897269397269397 }, "test_confusions": [ { "gold": "multi_intent_with_financial_risk", "predicted": "refund_missing_after_deadline", "count": 5 }, { "gold": "account_takeover", "predicted": "personal_data_or_payment_card_exposed", "count": 5 }, { "gold": "order_marked_delivered_but_not_received", "predicted": "multi_intent_with_financial_risk", "count": 5 }, { "gold": "change_delivery_address_after_dispatch", "predicted": "cancel_order_after_dispatch", "count": 5 }, { "gold": "fraud_or_phishing", "predicted": "account_takeover", "count": 3 }, { "gold": "missing_item_in_delivered_order", "predicted": "refund_amount_dispute", "count": 3 }, { "gold": "fraud_or_phishing", "predicted": "privacy_data_deletion_request", "count": 2 } ], "history": [ { "epoch": 1, "train_loss": 3.0446458036249333, "validation_loss": 2.9481882918964732, "validation_accuracy": 0.3181818181818182, "validation_macro_f1": 0.275695272469466 }, { "epoch": 2, "train_loss": 2.5696960145776924, "validation_loss": 2.4496376210992987, "validation_accuracy": 0.5818181818181818, "validation_macro_f1": 0.5069444444444445 }, { "epoch": 3, "train_loss": 1.8286097873340954, "validation_loss": 1.8624537467956543, "validation_accuracy": 0.7909090909090909, "validation_macro_f1": 0.7550946528219256 }, { "epoch": 4, "train_loss": 1.237558323686773, "validation_loss": 1.5112867398695513, "validation_accuracy": 0.8, "validation_macro_f1": 0.7386363636363636 }, { "epoch": 5, "train_loss": 0.899948058345101, "validation_loss": 1.34841853055087, "validation_accuracy": 0.8272727272727273, "validation_macro_f1": 0.7803030303030304 }, { "epoch": 6, "train_loss": 0.7283082962036133, "validation_loss": 1.267178336056796, "validation_accuracy": 0.8363636363636363, "validation_macro_f1": 0.7900596877869606 }, { "epoch": 7, "train_loss": 0.6527352463115346, "validation_loss": 1.2407976128838278, "validation_accuracy": 0.8090909090909091, "validation_macro_f1": 0.7660697887970614 }, { "epoch": 8, "train_loss": 0.6315497528422963, "validation_loss": 1.2309001250700518, "validation_accuracy": 0.8181818181818182, "validation_macro_f1": 0.7732847960120687 } ] }, "llm": { "task": "llm", "best_epoch": 4, "training_seconds": 87.89807533397106, "split_sizes": { "train": 800, "validation": 100, "test": 100 }, "validation": { "loss": 1.663049850463867, "accuracy": 0.71, "macro_f1": 0.6540935672514621 }, "test": { "loss": 1.4101154947280883, "accuracy": 0.8, "macro_f1": 0.754532967032967 }, "test_confusions": [ { "gold": "possible_mass_incident", "predicted": "damaged_item_first_steps", "count": 5 }, { "gold": "payment_pending_explanation", "predicted": "loyalty_points_missing", "count": 5 }, { "gold": "delivery_delay", "predicted": "change_delivery_address_before_dispatch", "count": 5 }, { "gold": "otp_code_not_received", "predicted": "promo_code_not_applied", "count": 3 }, { "gold": "refund_status_before_deadline", "predicted": "payment_pending_explanation", "count": 2 } ], "history": [ { "epoch": 1, "train_loss": 2.9181075382232664, "validation_loss": 2.8307441329956053, "validation_accuracy": 0.4, "validation_macro_f1": 0.3274270868020868 }, { "epoch": 2, "train_loss": 2.467692804336548, "validation_loss": 2.37255952835083, "validation_accuracy": 0.63, "validation_macro_f1": 0.5793831168831168 }, { "epoch": 3, "train_loss": 1.7932574939727783, "validation_loss": 1.940031542778015, "validation_accuracy": 0.7, "validation_macro_f1": 0.6371212121212121 }, { "epoch": 4, "train_loss": 1.2383545541763306, "validation_loss": 1.663049850463867, "validation_accuracy": 0.71, "validation_macro_f1": 0.6540935672514621 }, { "epoch": 5, "train_loss": 0.9075340437889099, "validation_loss": 1.513033676147461, "validation_accuracy": 0.71, "validation_macro_f1": 0.6506944444444445 }, { "epoch": 6, "train_loss": 0.7291339612007142, "validation_loss": 1.4201061391830445, "validation_accuracy": 0.71, "validation_macro_f1": 0.6527777777777778 }, { "epoch": 7, "train_loss": 0.6559483313560486, "validation_loss": 1.3904729557037354, "validation_accuracy": 0.71, "validation_macro_f1": 0.6506944444444445 } ] }, "auto": { "task": "auto", "best_epoch": 3, "training_seconds": 68.58642412500922, "split_sizes": { "train": 720, "validation": 90, "test": 90 }, "validation": { "loss": 1.5290983120600383, "accuracy": 0.8888888888888888, "macro_f1": 0.8611111111111112 }, "test": { "loss": 1.6726625601450602, "accuracy": 0.7777777777777778, "macro_f1": 0.712962962962963 }, "test_confusions": [ { "gold": "notification_settings", "predicted": "product_availability_notification", "count": 5 }, { "gold": "subscription_features", "predicted": "general_refund_timeline", "count": 5 }, { "gold": "how_to_contact_support", "predicted": "app_update_instructions", "count": 5 }, { "gold": "password_reset_instructions", "predicted": "product_availability_notification", "count": 5 } ], "history": [ { "epoch": 1, "train_loss": 2.8345796267191568, "validation_loss": 2.685704141192966, "validation_accuracy": 0.4222222222222222, "validation_macro_f1": 0.36838347901566293 }, { "epoch": 2, "train_loss": 2.3430909898546006, "validation_loss": 2.083689069747925, "validation_accuracy": 0.7333333333333333, "validation_macro_f1": 0.6763407072230602 }, { "epoch": 3, "train_loss": 1.6193762381871541, "validation_loss": 1.5290983120600383, "validation_accuracy": 0.8888888888888888, "validation_macro_f1": 0.8611111111111112 }, { "epoch": 4, "train_loss": 1.0734790404637655, "validation_loss": 1.1841111924913195, "validation_accuracy": 0.8888888888888888, "validation_macro_f1": 0.8611111111111112 }, { "epoch": 5, "train_loss": 0.7524434738688999, "validation_loss": 1.0104022343953452, "validation_accuracy": 0.8888888888888888, "validation_macro_f1": 0.8611111111111112 }, { "epoch": 6, "train_loss": 0.6061955703629388, "validation_loss": 0.9299269636472066, "validation_accuracy": 0.8888888888888888, "validation_macro_f1": 0.8611111111111112 } ] } }