davanstrien/aud-qwen3.5-4b-20260428
收藏官方服务:
资源简介:
这是一个由大语言模型(LLM)标注的数据集,通过classify-and-augment工具生成。数据集使用Qwen/Qwen3.5-4B模型进行标注,包含两个标签:positive和negative。原始数据共有180行,输出数据也是180行。标签分布显示,negative标签有178个,positive标签有2个,且没有合成数据。合成审计部分显示,positive类需要98个,生成了160个,验证了1个,保留了0个,接受率为0.6%。
This is an LLM-annotated dataset produced by classify-and-augment. The dataset was annotated using the Qwen/Qwen3.5-4B model and includes two labels: positive and negative. There are 180 input rows and 180 output rows. The label distribution shows 178 negative labels and 2 positive labels, with no synthetic data. The synthesis audit indicates that for the positive class, 98 were needed, 160 were generated, 1 was validated, 0 were kept, with an acceptance rate of 0.6%.
提供机构:
davanstrien


