open-llm-leaderboard/details_facebook__opt-66b
收藏Hugging Face2023-12-03 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_facebook__opt-66b
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facebook/opt-66b模型运行期间自动创建的,用于Open LLM Leaderboard。数据集包含122个配置,每个配置对应一个评估任务。数据集由5次运行创建,每次运行都有特定的分割,命名使用运行的时间戳。还有一个名为results的配置,存储所有运行的聚合结果,用于计算和显示Leaderboard上的聚合指标。
该数据集是在评估facebook/opt-66b模型运行期间自动创建的,用于Open LLM Leaderboard。数据集包含122个配置,每个配置对应一个评估任务。数据集由5次运行创建,每次运行都有特定的分割,命名使用运行的时间戳。还有一个名为results的配置,存储所有运行的聚合结果,用于计算和显示Leaderboard上的聚合指标。
提供机构:
open-llm-leaderboard
原始信息汇总
数据集概述
数据集简介
该数据集是在对模型 facebook/opt-66b 进行评估运行期间自动创建的,用于 Open LLM Leaderboard。
数据集结构
- 配置数量:122个配置,每个配置对应一个评估任务。
- 运行次数:数据集从5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片找到,分片名称使用运行的时间戳。
- 最新结果:"train" 分片总是指向最新的结果。
- 汇总结果:一个额外的配置 "results" 存储所有运行的汇总结果,用于计算和显示 Open LLM Leaderboard 上的汇总指标。
数据加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_facebook__opt-66b", "harness_gsm8k_5", split="train")
最新结果
以下是 2023-12-03T00:30:57.404111 运行的最新结果: python { "all": { "acc": 0.016679302501895376, "acc_stderr": 0.0035275958887224556 }, "harness|gsm8k|5": { "acc": 0.016679302501895376, "acc_stderr": 0.0035275958887224556 } }
配置详情
-
harness_arc_challenge_25
- 分片:2023_08_23T18_07_59.118983
- 路径:
**/details_harness|arc:challenge|25_2023-08-23T18:07:59.118983.parquet
- 路径:
- 分片:2023_08_24T00_29_23.220857
- 路径:
**/details_harness|arc:challenge|25_2023-08-24T00:29:23.220857.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|arc:challenge|25_2023-08-24T00:29:23.220857.parquet
- 路径:
- 分片:2023_08_23T18_07_59.118983
-
harness_drop_3
- 分片:2023_09_09T17_37_15.988083
- 路径:
**/details_harness|drop|3_2023-09-09T17-37-15.988083.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|drop|3_2023-09-09T17-37-15.988083.parquet
- 路径:
- 分片:2023_09_09T17_37_15.988083
-
harness_gsm8k_5
- 分片:2023_09_09T17_37_15.988083
- 路径:
**/details_harness|gsm8k|5_2023-09-09T17-37-15.988083.parquet
- 路径:
- 分片:2023_12_03T00_30_57.404111
- 路径:
**/details_harness|gsm8k|5_2023-12-03T00-30-57.404111.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|gsm8k|5_2023-12-03T00-30-57.404111.parquet
- 路径:
- 分片:2023_09_09T17_37_15.988083
-
harness_hellaswag_10
- 分片:2023_08_23T18_07_59.118983
- 路径:
**/details_harness|hellaswag|10_2023-08-23T18:07:59.118983.parquet
- 路径:
- 分片:2023_08_24T00_29_23.220857
- 路径:
**/details_harness|hellaswag|10_2023-08-24T00:29:23.220857.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|hellaswag|10_2023-08-24T00:29:23.220857.parquet
- 路径:
- 分片:2023_08_23T18_07_59.118983
-
harness_hendrycksTest_5
- 分片:2023_08_23T18_07_59.118983
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-formal_logic|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-global_facts|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-high_school_biology|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-high_school_geography|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-high_school_physics|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-high_school_psychology|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-high_school_statistics|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-high_school_us_history|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-high_school_world_history|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-human_aging|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-human_sexuality|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-international_law|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-jurisprudence|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-logical_fallacies|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-machine_learning|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-management|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-marketing|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-medical_genetics|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-miscellaneous|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-moral_disputes|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-moral_scenarios|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-nutrition|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-philosophy|5_2023-08-23T18:07:59.118983.parquet - 路径:
**/details_harness|hendrycksTest-prehistory|5_2023-08-23T18:07:59.118983.parquet - 路径:`**/details_harness|hendrycksTest-professional_accounting|5_2023-08-23T18:07
- 路径:
- 分片:2023_08_23T18_07_59.118983



