ClarusC64/clinical-drug-toxicity-instability-v0.1

该数据集用于评估模型是否能检测由于药物负荷超过清除能力引起的不稳定性。每个数据行代表一个简化的药物代谢场景，观察三个时间点。任务目标是确定药物调节是否保持稳定或趋向毒性不稳定。数据集的核心思想是当药物积累超过代谢清除能力时会发生药物毒性。不稳定性出现在药物水平快速上升、肝脏或肾脏清除能力下降、镇静或生理抑制增加、药物相互作用放大药理负荷或干预过晚时。数据集测试了这些信号之间的相互作用推理。预测目标为label=1表示药物毒性不稳定，label=0表示稳定的药理调节。

This dataset evaluates whether models can detect instability caused by pharmacological load exceeding clearance capacity. Each row represents a simplified drug metabolism scenario observed across three time points. The task is to determine whether pharmacological regulation remains stable or is moving toward toxic instability. The core stability idea is that drug toxicity occurs when drug accumulation exceeds metabolic clearance capacity. Instability emerges when drug levels rise rapidly, liver or renal clearance declines, sedation or physiological suppression increases, drug interactions amplify pharmacologic load, or intervention occurs too late. The dataset tests interaction reasoning across these signals. The prediction target is label = 1 for drug toxicity instability and label = 0 for stable pharmacologic regulation.