Sidsidney/Llama-Nemotron-Post-Training-Dataset

该数据集是SFT（监督微调）和RL（强化学习）数据的汇编，旨在提升原始Llama指导模型在数学、代码、一般推理和指令跟随方面的能力，支持NVIDIA发布的Llama-3.1-Nemotron-Ultra-253B-v1、Llama-3.3-Nemotron-Super-49B-v1和Llama-3.1-Nemotron-Nano-8B-v1模型。数据集包含数学、代码、科学、指令跟随、聊天和安全等多个类别的数据，通过公开和开放的语料库或合成生成的方式获取提示，并由多种模型生成响应。数据集主要用于社区继续改进开放模型，可自由用于训练和评估。

This dataset is a compilation of SFT and RL data that supports improvements of math, code, general reasoning, and instruction following capabilities of the original Llama instruct model, in support of NVIDIA’s release of Llama-3.1-Nemotron-Ultra-253B-v1, Llama-3.3-Nemotron-Super-49B-v1 and Llama-3.1-Nemotron-Nano-8B-v1. The dataset includes data categories such as math, code, science, instruction following, chat, and safety. Prompts have been sourced from either public and open corpus or synthetically generated, and responses were synthetically generated by a variety of models. The dataset is intended to be used by the community to continue to improve open models and may be freely used to train and evaluate.