davanstrien/imdb-classify-augment-v5-distfilter

这是一个由LLM（大型语言模型）标注的IMDB电影评论数据集，使用classify-and-augment工具生成。数据集包含正面（positive）和负面（negative）两类情感标签，使用HuggingFaceTB/SmolLM3-3B模型处理。原始输入20行数据，最终输出25行数据（包含真实数据和合成数据）。标签分布显示：负面标签13个（9真实+4合成），正面标签12个（11真实+1合成）。合成数据经过模型自一致性检查验证，接受率为50%。This is an LLM-annotated IMDB movie review dataset produced using the classify-and-augment tool. The dataset contains two sentiment labels: positive and negative, processed by the HuggingFaceTB/SmolLM3-3B model. It started with 20 input rows and produced 25 output rows (including real and synthetic data). Label distribution shows: 13 negative labels (9 real + 4 synthetic) and 12 positive labels (11 real + 1 synthetic). Synthetic data was validated through model self-consistency checks with a 50% acceptance rate.