jackf857/llama-3-8b-base-margin-dpo-hh-harmless-margin-log-beta0.01

该数据集是从margin-DPO训练运行中导出的每步边缘摘要统计信息，包含661行训练数据。数据集来源于模型仓库jackf857/llama-3-8b-base-margin-dpo-hh-harmless-beta0.01，基础模型为W-61/llama-3-8b-base-sft-hh-harmless-4xh200。数据集包含多个统计列，如epoch、step、batch_size、mean、std等，以及每步边缘的详细样本信息。数据集混合器配置为100%使用Anthropic/hh-rlhf数据。

This dataset contains per-step margin summary statistics exported from a margin-DPO training run, with 661 rows of training data. The dataset originates from the model repository jackf857/llama-3-8b-base-margin-dpo-hh-harmless-beta0.01, with the base model being W-61/llama-3-8b-base-sft-hh-harmless-4xh200. The dataset includes various statistical columns such as epoch, step, batch_size, mean, std, etc., as well as detailed sample information for each steps margins. The dataset mixer is configured to use 100% Anthropic/hh-rlhf data.