open-llm-leaderboard-old/details_Plaban81__Moe-4x7b-math-reason-code
收藏Hugging Face2024-02-01 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_Plaban81__Moe-4x7b-math-reason-code
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Plaban81/Moe-4x7b-math-reason-code时自动创建的,主要用于在Open LLM Leaderboard上展示评估结果。数据集包含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,并用于计算和展示Open LLM Leaderboard上的聚合指标。
该数据集是在评估模型Plaban81/Moe-4x7b-math-reason-code时自动创建的,主要用于在Open LLM Leaderboard上展示评估结果。数据集包含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,并用于计算和展示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
数据集创建
- 创建背景:该数据集是在评估模型Plaban81/Moe-4x7b-math-reason-code在Open LLM Leaderboard上的运行过程中自动创建的。
数据集结构
- 配置数量: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,每个运行结果作为一个特定的分割(split)存储在每个配置中,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。
- 结果汇总:一个额外的配置"results"存储所有运行的汇总结果,用于计算和显示在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_Plaban81__Moe-4x7b-math-reason-code", "harness_winogrande_5", split="train")
最新结果
- 最新结果时间戳:2024-02-01T23:02:06.191650
- 结果示例:
python
{
"all": {
"acc": 0.61366513110882,
"acc_stderr": 0.033073367558772764,
"acc_norm": 0.6160976239521233,
"acc_norm_stderr": 0.03373653941415801,
"mc1": 0.4039167686658507,
"mc1_stderr": 0.017177276822584284,
"mc2": 0.5611833547243651,
"mc2_stderr": 0.015990413066061377
},
"harness|arc:challenge|25": {
"acc": 0.5887372013651877,
"acc_stderr": 0.014379441068522084,
"acc_norm": 0.6254266211604096,
"acc_norm_stderr": 0.014144193471893454
},
"harness|hellaswag|10": {
"acc": 0.652459669388568,
"acc_stderr": 0.004752158936871871,
"acc_norm": 0.8386775542720574,
"acc_norm_stderr": 0.0036707636737929633
},
其他任务结果省略
}
配置详情
-
配置名称:harness_arc_challenge_25
- 数据文件:
- 分割:2024_02_01T23_02_06.191650
- 路径:
**/details_harness|arc:challenge|25_2024-02-01T23-02-06.191650.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|arc:challenge|25_2024-02-01T23-02-06.191650.parquet
- 路径:
- 分割:2024_02_01T23_02_06.191650
- 数据文件:
-
配置名称:harness_gsm8k_5
- 数据文件:
- 分割:2024_02_01T23_02_06.191650
- 路径:
**/details_harness|gsm8k|5_2024-02-01T23-02-06.191650.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|gsm8k|5_2024-02-01T23-02-06.191650.parquet
- 路径:
- 分割:2024_02_01T23_02_06.191650
- 数据文件:
-
配置名称:harness_hellaswag_10
- 数据文件:
- 分割:2024_02_01T23_02_06.191650
- 路径:
**/details_harness|hellaswag|10_2024-02-01T23-02-06.191650.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|hellaswag|10_2024-02-01T23-02-06.191650.parquet
- 路径:
- 分割:2024_02_01T23_02_06.191650
- 数据文件:
-
配置名称:harness_hendrycksTest_5
- 数据文件:
- 分割:2024_02_01T23_02_06.191650
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-anatomy|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-astronomy|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-business_ethics|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-college_biology|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-college_chemistry|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-college_computer_science|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-college_mathematics|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-college_medicine|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-college_physics|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-computer_security|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-econometrics|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-formal_logic|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-global_facts|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-high_school_biology|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-high_school_chemistry|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-high_school_computer_science|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-high_school_european_history|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-high_school_geography|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-high_school_government_and_politics|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-high_school_macroeconomics|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-high_school_mathematics|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-high_school_microeconomics|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-high_school_physics|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-high_school_psychology|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-high_school_statistics|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-high_school_us_history|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-high_school_world_history|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-human_aging|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-human_sexuality|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-international_law|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-jurisprudence|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-logical_fallacies|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-machine_learning|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-management|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-marketing|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-medical_genetics|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-miscellaneous|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-moral_disputes|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-moral_scenarios|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-nutrition|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-philosophy|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-prehistory|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-professional_accounting|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-professional_law|5_2024-02-01T23-02-06.191650.parquet**/details_harness|hendrycksTest-professional_medicine|5_2024-02-01T23-02-06.191650.parquet- `**/details_harness|hendrycksTest-professional_psychology|5_2024-02-01T23-02-06.1
- 路径:
- 分割:2024_02_01T23_02_06.191650
- 数据文件:



