open-llm-leaderboard/details_heegyu__LIMA-13b-hf
收藏Hugging Face2023-10-22 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_heegyu__LIMA-13b-hf
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是在模型heegyu/LIMA-13b-hf在Open LLM Leaderboard上的评估运行期间自动创建的。数据集由64个配置组成,每个配置对应一个评估任务。数据集由2次运行创建,每次运行可以在每个配置中找到特定的分割,分割名称使用运行的时间戳命名。train分割始终指向最新的结果。此外,一个名为results的配置存储了所有运行的聚合结果,并用于计算和显示Open LLM Leaderboard上的聚合指标。
该数据集是在模型heegyu/LIMA-13b-hf在Open LLM Leaderboard上的评估运行期间自动创建的。数据集由64个配置组成,每个配置对应一个评估任务。数据集由2次运行创建,每次运行可以在每个配置中找到特定的分割,分割名称使用运行的时间戳命名。train分割始终指向最新的结果。此外,一个名为results的配置存储了所有运行的聚合结果,并用于计算和显示Open LLM 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
数据集名称
- 名称: Evaluation run of heegyu/LIMA-13b-hf
数据集来源
- 创建背景: 该数据集是在对模型 heegyu/LIMA-13b-hf 进行评估时自动创建的,评估结果发布在 Open LLM Leaderboard 上。
数据集结构
- 配置数量: 该数据集包含 64 个配置,每个配置对应一个评估任务。
- 数据分割: 每个配置包含多个分割,分割名称使用评估运行的时间戳。"train" 分割始终指向最新的结果。
- 结果汇总: 一个额外的配置 "results" 存储所有评估运行的汇总结果,用于计算和显示在 Open LLM Leaderboard 上的聚合指标。
数据加载示例
- Python 代码示例: python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_heegyu__LIMA-13b-hf", "harness_winogrande_5", split="train")
最新结果
- 最新结果链接: 最新结果来自 run 2023-10-22T00:08:39.312434
- 结果示例: python { "all": { "em": 0.0020973154362416107, "em_stderr": 0.000468506503036833, "f1": 0.05783347315436248, "f1_stderr": 0.0013197558360646307, "acc": 0.4303028471618438, "acc_stderr": 0.009812237277361156 }, "harness|drop|3": { "em": 0.0020973154362416107, "em_stderr": 0.000468506503036833, "f1": 0.05783347315436248, "f1_stderr": 0.0013197558360646307 }, "harness|gsm8k|5": { "acc": 0.0887035633055345, "acc_stderr": 0.00783145873705871 }, "harness|winogrande|5": { "acc": 0.7719021310181531, "acc_stderr": 0.0117930158176636 } }
配置详情
-
配置名称: harness_arc_challenge_25
- 数据文件:
- 分割: 2023_08_17T19_40_51.725558
- 路径: **/details_harness|arc:challenge|25_2023-08-17T19:40:51.725558.parquet
- 分割: latest
- 路径: **/details_harness|arc:challenge|25_2023-08-17T19:40:51.725558.parquet
- 分割: 2023_08_17T19_40_51.725558
- 数据文件:
-
配置名称: harness_drop_3
- 数据文件:
- 分割: 2023_10_22T00_08_39.312434
- 路径: **/details_harness|drop|3_2023-10-22T00-08-39.312434.parquet
- 分割: latest
- 路径: **/details_harness|drop|3_2023-10-22T00-08-39.312434.parquet
- 分割: 2023_10_22T00_08_39.312434
- 数据文件:
-
配置名称: harness_gsm8k_5
- 数据文件:
- 分割: 2023_10_22T00_08_39.312434
- 路径: **/details_harness|gsm8k|5_2023-10-22T00-08-39.312434.parquet
- 分割: latest
- 路径: **/details_harness|gsm8k|5_2023-10-22T00-08-39.312434.parquet
- 分割: 2023_10_22T00_08_39.312434
- 数据文件:
-
配置名称: harness_hellaswag_10
- 数据文件:
- 分割: 2023_08_17T19_40_51.725558
- 路径: **/details_harness|hellaswag|10_2023-08-17T19:40:51.725558.parquet
- 分割: latest
- 路径: **/details_harness|hellaswag|10_2023-08-17T19:40:51.725558.parquet
- 分割: 2023_08_17T19_40_51.725558
- 数据文件:
-
配置名称: harness_hendrycksTest_5
- 数据文件:
- 分割: 2023_08_17T19_40_51.725558
- 路径:
- **/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-anatomy|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-astronomy|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-business_ethics|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-college_biology|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-college_chemistry|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-college_computer_science|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-college_mathematics|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-college_medicine|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-college_physics|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-computer_security|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-conceptual_physics|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-econometrics|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-electrical_engineering|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-formal_logic|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-global_facts|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-high_school_biology|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-high_school_european_history|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-high_school_geography|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-high_school_physics|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-high_school_psychology|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-high_school_statistics|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-high_school_us_history|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-high_school_world_history|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-human_aging|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-human_sexuality|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-international_law|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-jurisprudence|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-logical_fallacies|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-machine_learning|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-management|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-marketing|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-medical_genetics|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-miscellaneous|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-moral_disputes|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-moral_scenarios|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-nutrition|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-philosophy|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-prehistory|5_2023-08-17T19:40:51.725558.parquet
- **/details_harness|hendrycksTest-professional
- 路径:
- 分割: 2023_08_17T19_40_51.725558
- 数据文件:



