open-llm-leaderboard-old/details_allknowingroger__JupiterMerge-7B-slerp
收藏Hugging Face2024-04-10 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_allknowingroger__JupiterMerge-7B-slerp
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是在评估模型allknowingroger/JupiterMerge-7B-slerp时自动创建的,包含63个配置,每个配置对应一个评估任务。数据集由1次运行生成,每次运行的结果存储为特定配置中的一个分割,分割名称使用运行的时间戳。train分割始终指向最新的结果。此外,还有一个名为results的配置,存储了所有运行的聚合结果,用于计算和展示在Open LLM Leaderboard上的聚合指标。
该数据集是在评估模型allknowingroger/JupiterMerge-7B-slerp时自动创建的,包含63个配置,每个配置对应一个评估任务。数据集由1次运行生成,每次运行的结果存储为特定配置中的一个分割,分割名称使用运行的时间戳。train分割始终指向最新的结果。此外,还有一个名为results的配置,存储了所有运行的聚合结果,用于计算和展示在Open LLM Leaderboard上的聚合指标。
提供机构:
open-llm-leaderboard-old
原始信息汇总
数据集概述
数据集基本信息
- 数据集名称: Evaluation run of allknowingroger/JupiterMerge-7B-slerp
- 数据集创建: 该数据集是在评估模型 allknowingroger/JupiterMerge-7B-slerp 在 Open LLM Leaderboard 上的自动创建的。
数据集组成
- 配置数量: 该数据集包含 63 个配置,每个配置对应一个评估任务。
- 数据来源: 数据集从 1 次运行中创建,每个运行可以在每个配置中作为一个特定的分割找到,分割名称使用运行的时间戳。"train" 分割总是指向最新的结果。
数据集加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_allknowingroger__JupiterMerge-7B-slerp", "harness_winogrande_5", split="train")
最新结果
- 最新结果时间戳: 2024-04-10T20:59:01.989344
- 结果概览: python { "all": { "acc": 0.6563205074584324, "acc_stderr": 0.03196275545073811, "acc_norm": 0.6555506629068194, "acc_norm_stderr": 0.03263347097242048, "mc1": 0.591187270501836, "mc1_stderr": 0.017209952151641724, "mc2": 0.7359824971805283, "mc2_stderr": 0.014519913019364611 }, "harness|arc:challenge|25": { "acc": 0.7141638225255973, "acc_stderr": 0.013203196088537372, "acc_norm": 0.7380546075085325, "acc_norm_stderr": 0.012849054826858107 }, "harness|hellaswag|10": { "acc": 0.7161919936267676, "acc_stderr": 0.004499233874427508, "acc_norm": 0.8892650866361282, "acc_norm_stderr": 0.0031316226281990827 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.36, "acc_stderr": 0.04824181513244218, "acc_norm": 0.36, "acc_norm_stderr": 0.04824181513244218 }, # 更多任务结果... }
配置详情
-
配置名称: harness_arc_challenge_25
- 数据文件:
- 分割: 2024_04_10T20_59_01.989344
- 路径: **/details_harness|arc:challenge|25_2024-04-10T20-59-01.989344.parquet
- 分割: latest
- 路径: **/details_harness|arc:challenge|25_2024-04-10T20-59-01.989344.parquet
- 分割: 2024_04_10T20_59_01.989344
- 数据文件:
-
配置名称: harness_gsm8k_5
- 数据文件:
- 分割: 2024_04_10T20_59_01.989344
- 路径: **/details_harness|gsm8k|5_2024-04-10T20-59-01.989344.parquet
- 分割: latest
- 路径: **/details_harness|gsm8k|5_2024-04-10T20-59-01.989344.parquet
- 分割: 2024_04_10T20_59_01.989344
- 数据文件:
-
配置名称: harness_hellaswag_10
- 数据文件:
- 分割: 2024_04_10T20_59_01.989344
- 路径: **/details_harness|hellaswag|10_2024-04-10T20-59-01.989344.parquet
- 分割: latest
- 路径: **/details_harness|hellaswag|10_2024-04-10T20-59-01.989344.parquet
- 分割: 2024_04_10T20_59_01.989344
- 数据文件:
-
配置名称: harness_hendrycksTest_5
- 数据文件:
- 分割: 2024_04_10T20_59_01.989344
- 路径:
- **/details_harness|hendrycksTest-abstract_algebra|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-anatomy|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-astronomy|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-business_ethics|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-clinical_knowledge|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-college_biology|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-college_chemistry|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-college_computer_science|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-college_mathematics|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-college_medicine|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-college_physics|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-computer_security|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-conceptual_physics|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-econometrics|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-electrical_engineering|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-elementary_mathematics|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-formal_logic|5_2024-04-10T20-59-01.989344.parquet
- **/details_harness|hendrycksTest-global_facts|5_2024-04-10T20-59-01.989344.parquet
更多路径...
- 路径:
- 分割: 2024_04_10T20_59_01.989344
- 数据文件:



