akhilpandey95/scienceeval
收藏Hugging Face2026-04-29 更新2026-05-03 收录
下载链接：
https://hf-mirror.com/datasets/akhilpandey95/scienceeval
下载链接
链接失效反馈官方服务：
资源简介：
scienceeval数据集是一个用于评估模型在广泛科学基准测试中能力的工具。它包含69,315个科学基准项目的全语料库基准嵌入和投影坐标。嵌入是通过`pplx-embed-v1-0.6B`生成，并使用PCA降维到50维，再通过t-SNE、UMAP和PCA进行投影。数据集还包括嵌入矩阵、元数据、投影坐标和质量指标等文件。
The scienceeval dataset is a tool for evaluating the capabilities of models across a broad swath of science benchmarks. It contains full-corpus benchmark embeddings and projection coordinates for 69,315 science benchmark items. The embeddings were generated with `pplx-embed-v1-0.6B`, reduced to 50 dimensions with PCA, and projected with t-SNE, UMAP, and PCA. The dataset also includes files such as embedding matrices, metadata, projection coordinates, and quality metrics.
提供机构：
akhilpandey95
