ermiaazarkhalili/orca-mini-v1-high-prob-qwen-0.5b-10k

该数据集包含来自[pankajmathur/orca_mini_v1_dataset](https://huggingface.co/datasets/pankajmathur/orca_mini_v1_dataset)的句子，其中模型`Qwen/Qwen2.5-0.5B`以≥90%的概率预测句号前的最后一个词。数据集详细记录了源数据集、使用的模型、概率阈值以及模式字段，如model_id、dataset_id、sentence_prefix、predicted_token、actual_token、probability和num_tokens。该数据集适用于文本生成任务，规模在10K到100K之间。

This dataset contains sentences from [pankajmathur/orca_mini_v1_dataset](https://huggingface.co/datasets/pankajmathur/orca_mini_v1_dataset) where the model `Qwen/Qwen2.5-0.5B` predicts the token before the final period with ≥90% probability. The dataset details the source dataset, model used, probability threshold, and schema fields such as model_id, dataset_id, sentence_prefix, predicted_token, actual_token, probability, and num_tokens. It is intended for text-generation tasks and has a size category of 10K<n<100K.