Evaluation Setting of Merging Large Language Models with Heterogeneous Architecture

This comparison table presents the experimental evaluation settings used for merging large language models with heterogeneous architectures, including the datasets, tasks, architecture, evaluation metrics, model combinations, and baseline methods employed to assess merging performance.

本对比表格展示了用于融合具有异构架构的大语言模型（Large Language Model）的实验评估设置，涵盖了评估该融合性能时所使用的数据集、任务、模型架构、评估指标、模型组合以及基准方法。