open-llm-leaderboard-old/details_lamhieu__ghost-7b-v0.9.1
收藏数据集概述
数据集简介
该数据集是在对模型 lamhieu/ghost-7b-v0.9.1 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:63个配置,每个配置对应一个评估任务。
- 数据来源:数据集由1次运行创建,每个运行在每个配置中作为一个特定的分割存在,分割名称使用运行的时间戳。
- 最新结果:"train" 分割始终指向最新的结果。
- 聚合结果:一个额外的配置 "results" 存储所有运行的聚合结果,用于计算和显示 Open LLM Leaderboard 上的聚合指标。
数据加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_lamhieu__ghost-7b-v0.9.1", "harness_winogrande_5", split="train")
最新结果
以下是 2024-02-22T12:21:15.645809 运行的最新结果:
python { "all": { "acc": 0.5461963762782457, "acc_stderr": 0.0341839674901544, "acc_norm": 0.5516542798358821, "acc_norm_stderr": 0.034909524738250194, "mc1": 0.29865361077111385, "mc1_stderr": 0.016021570613768542, "mc2": 0.43956086098918057, "mc2_stderr": 0.015308355019122989 }, "harness|arc:challenge|25": { "acc": 0.5170648464163823, "acc_stderr": 0.014602878388536595, "acc_norm": 0.5537542662116041, "acc_norm_stderr": 0.01452670554853998 }, "harness|hellaswag|10": { "acc": 0.5828520215096594, "acc_stderr": 0.004920800313232742, "acc_norm": 0.770264887472615, "acc_norm_stderr": 0.004198027272982672 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.25, "acc_stderr": 0.04351941398892446, "acc_norm": 0.25, "acc_norm_stderr": 0.04351941398892446 }, # 其他任务的结果... }
配置详情
-
config_name: harness_arc_challenge_25
- data_files:
- split: 2024_02_22T12_21_15.645809
**/details_harness|arc:challenge|25_2024-02-22T12-21-15.645809.parquet
- split: latest
**/details_harness|arc:challenge|25_2024-02-22T12-21-15.645809.parquet
- split: 2024_02_22T12_21_15.645809
- data_files:
-
config_name: harness_gsm8k_5
- data_files:
- split: 2024_02_22T12_21_15.645809
**/details_harness|gsm8k|5_2024-02-22T12-21-15.645809.parquet
- split: latest
**/details_harness|gsm8k|5_2024-02-22T12-21-15.645809.parquet
- split: 2024_02_22T12_21_15.645809
- data_files:
-
config_name: harness_hellaswag_10
- data_files:
- split: 2024_02_22T12_21_15.645809
**/details_harness|hellaswag|10_2024-02-22T12-21-15.645809.parquet
- split: latest
**/details_harness|hellaswag|10_2024-02-22T12-21-15.645809.parquet
- split: 2024_02_22T12_21_15.645809
- data_files:
-
config_name: harness_hendrycksTest_5
- data_files:
- split: 2024_02_22T12_21_15.645809
**/details_harness|hendrycksTest-abstract_algebra|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-anatomy|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-astronomy|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-business_ethics|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-college_biology|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-college_chemistry|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-college_computer_science|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-college_mathematics|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-college_medicine|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-college_physics|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-computer_security|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-econometrics|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-formal_logic|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-global_facts|5_2024-02-22T12-21-15.645809.parquet**/details_harness|hendrycksTest-high_school_biology|5_2024-02-22T12-21-15.645809.parquet
其他文件路径...
- split: 2024_02_22T12_21_15.645809
- data_files:
}



