open-llm-leaderboard/details_ehartford__WizardLM-1.0-Uncensored-Llama2-13b
收藏数据集概述
数据集简介
该数据集是在评估模型ehartford/WizardLM-1.0-Uncensored-Llama2-13b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:64个配置,每个配置对应一个评估任务。
- 运行次数:数据集由4次运行创建。每次运行在每个配置中作为一个特定的分片存在,分片名称使用运行的时间戳。
- 最新结果:"train"分片始终指向最新的结果。
- 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上的聚合指标。
数据加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_ehartford__WizardLM-1.0-Uncensored-Llama2-13b", "harness_winogrande_5", split="train")
最新结果
以下是2023-10-22T09:23:28.206908运行的最新结果: python { "all": { "em": 0.07403523489932885, "em_stderr": 0.0026813660805584437, "f1": 0.1393938758389259, "f1_stderr": 0.002927612388923708, "acc": 0.43689851379839195, "acc_stderr": 0.010827222471217795 }, "harness|drop|3": { "em": 0.07403523489932885, "em_stderr": 0.0026813660805584437, "f1": 0.1393938758389259, "f1_stderr": 0.002927612388923708 }, "harness|gsm8k|5": { "acc": 0.1326762699014405, "acc_stderr": 0.009343929131442217 }, "harness|winogrande|5": { "acc": 0.7411207576953434, "acc_stderr": 0.012310515810993372 } }
配置详情
-
harness_arc_challenge_25
- 分片:2023_08_09T13_52_58.129270
- 路径:
**/details_harness|arc:challenge|25_2023-08-09T13:52:58.129270.parquet
- 路径:
- 分片:2023_08_09T13_58_22.615807
- 路径:
**/details_harness|arc:challenge|25_2023-08-09T13:58:22.615807.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|arc:challenge|25_2023-08-09T13:58:22.615807.parquet
- 路径:
- 分片:2023_08_09T13_52_58.129270
-
harness_drop_3
- 分片:2023_10_21T18_02_33.416249
- 路径:
**/details_harness|drop|3_2023-10-21T18-02-33.416249.parquet
- 路径:
- 分片:2023_10_22T09_23_28.206908
- 路径:
**/details_harness|drop|3_2023-10-22T09-23-28.206908.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|drop|3_2023-10-22T09-23-28.206908.parquet
- 路径:
- 分片:2023_10_21T18_02_33.416249
-
harness_gsm8k_5
- 分片:2023_10_21T18_02_33.416249
- 路径:
**/details_harness|gsm8k|5_2023-10-21T18-02-33.416249.parquet
- 路径:
- 分片:2023_10_22T09_23_28.206908
- 路径:
**/details_harness|gsm8k|5_2023-10-22T09-23-28.206908.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|gsm8k|5_2023-10-22T09-23-28.206908.parquet
- 路径:
- 分片:2023_10_21T18_02_33.416249
-
harness_hellaswag_10
- 分片:2023_08_09T13_52_58.129270
- 路径:
**/details_harness|hellaswag|10_2023-08-09T13:52:58.129270.parquet
- 路径:
- 分片:2023_08_09T13_58_22.615807
- 路径:
**/details_harness|hellaswag|10_2023-08-09T13:58:22.615807.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|hellaswag|10_2023-08-09T13:58:22.615807.parquet
- 路径:
- 分片:2023_08_09T13_52_58.129270
-
harness_hendrycksTest_5
- 分片:2023_08_09T13_52_58.129270
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-formal_logic|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-global_facts|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-high_school_biology|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-high_school_geography|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-high_school_physics|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-high_school_psychology|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-high_school_statistics|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-high_school_us_history|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-high_school_world_history|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-human_aging|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-human_sexuality|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-international_law|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-jurisprudence|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-logical_fallacies|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-machine_learning|5_2023-08-09T13:52:58.129270.parquet - 路径:
**/details_harness|hendrycksTest-management|5_2023-08-09T13:52:58.129270.parquet - 路径:`**/details_harness|hendrycksTest-marketing|5_2023-08-09T13:52
- 路径:
- 分片:2023_08_09T13_52_58.129270



