Ramitha/alqa-results-mistral-lbert

该数据集包含多个字段，主要涉及问题、答案、片段以及与不同模型（如Mistral、Llama、Gemma、Falcon）相关的生成内容和相似性度量。数据集还包含了问题和答案的嵌入表示，以及不同模型生成的答案之间的相似性评分。此外，数据集还记录了问题重建误差和案例对齐反向生成的评分。这些字段表明该数据集可能用于评估不同模型在问答任务中的表现，以及问题和答案之间的语义相似性分析。

The dataset contains multiple features, such as questions, answers, snippets, and their corresponding embeddings and similarity scores. Specific features include questions, answers, snippets, answers from different models (e.g., answerMistral, question_answerMistral_llama, etc.), and their embeddings. Additionally, it includes some similarity scores (e.g., question_answer_similarity, question_snippet_similarity) and reconstruction errors (e.g., question_reconstruction_llama_error). The dataset is split into one partition named rawcases with 2084 samples.