bermaneh/pde-mc-logprob-results-v3-deepseek
收藏官方服务:
资源简介:
该数据集名为pde-mc-logprob-results-v3-deepseek,是使用DeepSeek-R1-Distill-Qwen-32B模型在v3数据集上进行的多项选择题(MC)文本提取结果。v3数据集包含128行数据、8种条件和9种问题类型,总共有1152行和19列。数据集记录了模型在多项选择题上的预测结果,包括候选答案、对数概率、正确性等指标,用于评估模型性能。
The dataset is named pde-mc-logprob-results-v3-deepseek and contains results from DeepSeek-R1-Distill-Qwen-32B models multiple-choice (MC) text-extraction on the v3 dataset, which includes 128 rows, 8 conditions, and 9 question types. It consists of 1152 rows and 19 columns, capturing model predictions such as candidate answers, log probabilities, correctness, and other metrics for performance evaluation.
提供机构:
bermaneh


