hamishivi/tulu-2-unfiltered

Tulu 2 Unfiltered是一个未经过滤的版本的数据集，包含了FLAN、Open Assistant 1、ShareGPT、GPT4-Alpaca、Code-Alpaca、LIMA、WizardLM Evol Instruct、Open-Orca等多种数据源的混合，以及科学文档理解任务的例子，如问答、事实核查、总结和信息抽取等。Tulu 2 Unfiltered is an unfiltered version of the dataset, which includes a mix of sources such as FLAN, Open Assistant 1, ShareGPT, GPT4-Alpaca, Code-Alpaca, LIMA, WizardLM Evol Instruct, Open-Orca, and examples of scientific document understanding tasks like question answering, fact-checking, summarization, and information extraction.