llm-jp/aya-ja-evol-inst
收藏Hugging Face2025-02-04 更新2025-02-15 收录
下载链接:
https://hf-mirror.com/datasets/llm-jp/aya-ja-evol-inst
下载链接
链接失效反馈官方服务:
资源简介:
aya-ja-eval-inst是一个由LLM-jp项目创建的日语偏好数据集,用于文本生成任务。数据集包含了通过Qwen/Qwen2.5-32B-Instruct模型生成的选择的回应和通过llm-jp/llm-jp-3-1.8b-instruct模型生成的被拒绝的回应。数据集规模在10K到100K之间,是合成的数据。
aya-ja-eval-inst is a Japanese preference dataset created by the LLM-jp project for text generation tasks. The dataset consists of selected responses generated by the Qwen/Qwen2.5-32B-Instruct model and rejected responses generated by the llm-jp/llm-jp-3-1.8b-instruct model. The dataset size is between 10K and 100K and is synthetic data.
提供机构:
llm-jp


