W-61/llama-3-8b-base-new-dpo-hh-harmless-4xh200-batch-64-q_t-0.45-s_star-0.4-eta-1-margin-log

该数据集是从New-DPO训练运行中导出的每步边缘摘要统计信息。包含训练过程中的多个特征，如epoch、step、batch_size、mean、std、min、p10、median、p90、max、pos_frac、sample和npy等。数据集的来源运行包括模型仓库ID、基础模型、训练运行名称、W&B项目、训练器类型、边缘日志路径等详细信息。此外，还提供了边缘训练参数和数据集混合器的信息。

Per-step margin summary statistics exported from a New-DPO training run. The dataset includes multiple features during the training process, such as epoch, step, batch_size, mean, std, min, p10, median, p90, max, pos_frac, sample, and npy. The source run of the dataset includes details such as model repo id, base model, training run name, W&B project, trainer type, margin log path, etc. Additionally, margin training arguments and dataset mixer information are provided.