maanas-writer/mem_agent-model_based-rl-memoryagent-14b-barexamqa-train-c27000-t128-10s-agnostic

该数据集包含了问题、上下文、正确答案、响应、提取的答案、最终内存、内存长度、响应长度、上下文长度、相似度、是否正确、响应URL、响应URL索引、模型名称、是否截断、原始上下文长度、截断上下文长度和移除的标记数等字段。数据集分为训练集，提供了字节数和示例数。同时，包含了一个默认配置，指定了训练数据的路径。

The dataset includes fields such as question, context, ground truth, response, extracted answer, final memory, memory length, response length, context length, similarity, correctness, response URL, response URL index, model name, whether truncated, original context length, truncated context length, and number of tokens removed. The dataset is split into a training set with provided byte size and number of examples. It also contains a default configuration specifying the path to the training data.