NLTF-mock/tw-instruct-500k-Q-R1

台灣常見任務對話集是一個合成資料集，由參考文本和常見種子提示生成，旨在讓語言模型了解和生成台灣社會的對話。資料集包含參考文本生成的指令對話集和常見種子提示生成的台灣場景對話集。這個資料集可用於訓練語言模型，使其具有一般繁體中文的對話能力。

The Common Task-Oriented Dialogues in Taiwan dataset is a synthetic dataset composed of reference-based and reference-free subsets, aiming to train language models to understand and generate dialogue in the context of Taiwan society. The dataset includes dialogue sets generated from reference texts and Taiwan scene dialogue sets generated from common seed prompts. This dataset can be used to train language models to have general Chinese conversation capabilities.