Eliciting-Contexts/jailbreaking

该数据集包含模板（template）、变量文本（variable_text）、期望文本（desired_text）、非期望文本（undesired_text）、故事类型（story_type）、人工答案（human_answer）、预测单词（predicted_word）以及期望和非期望文本的详细信息（desired_details和undesired_details）。详细信息包括logit值、排名和单词。数据集分为测试集（test），共有20个示例，大小为9896字节。The dataset includes fields such as template, variable_text, desired_text, undesired_text, story_type, human_answer, predicted_word, and detailed information of desired and undesired texts (desired_details and undesired_details). The detailed information consists of logit values, ranks, and words. The dataset is split into a test set with 20 examples, totaling 9896 bytes.