open-llm-leaderboard-old/details_Ppoyaa__Alpha-Mistral-7B-Instruct
收藏数据集概述
数据集简介
该数据集是在评估模型Ppoyaa/Alpha-Mistral-7B-Instruct在Open LLM Leaderboard上的运行过程中自动创建的。数据集包含63个配置,每个配置对应一个评估任务。
数据集结构
数据集由1次运行创建,每个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。"train"分割始终指向最新的结果。
额外配置
一个额外的配置"results"存储了所有运行的聚合结果,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_Ppoyaa__Alpha-Mistral-7B-Instruct", "harness_winogrande_5", split="train")
最新结果
以下是最新结果(来自2024-04-15T11:01:11.453575运行)的摘要:
python { "all": { "acc": 0.5292991804932395, "acc_stderr": 0.03414202331308852, "acc_norm": 0.5333047977390053, "acc_norm_stderr": 0.03486011046426882, "mc1": 0.3635250917992656, "mc1_stderr": 0.016838862883965827, "mc2": 0.5296347812834613, "mc2_stderr": 0.015236773829187759 }, "harness|arc:challenge|25": { "acc": 0.5878839590443686, "acc_stderr": 0.014383915302225403, "acc_norm": 0.636518771331058, "acc_norm_stderr": 0.014056207319068282 }, "harness|hellaswag|10": { "acc": 0.6070503883688508, "acc_stderr": 0.004874076250521577, "acc_norm": 0.8149770961959769, "acc_norm_stderr": 0.0038752253693657315 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.27, "acc_stderr": 0.044619604333847394, "acc_norm": 0.27, "acc_norm_stderr": 0.044619604333847394 }, # 其他任务的结果... }
配置详情
-
config_name: harness_arc_challenge_25
- 分割: 2024_04_15T11_01_11.453575
- 路径: **/details_harness|arc:challenge|25_2024-04-15T11-01-11.453575.parquet
- 分割: latest
- 路径: **/details_harness|arc:challenge|25_2024-04-15T11-01-11.453575.parquet
- 分割: 2024_04_15T11_01_11.453575
-
config_name: harness_gsm8k_5
- 分割: 2024_04_15T11_01_11.453575
- 路径: **/details_harness|gsm8k|5_2024-04-15T11-01-11.453575.parquet
- 分割: latest
- 路径: **/details_harness|gsm8k|5_2024-04-15T11-01-11.453575.parquet
- 分割: 2024_04_15T11_01_11.453575
-
config_name: harness_hellaswag_10
- 分割: 2024_04_15T11_01_11.453575
- 路径: **/details_harness|hellaswag|10_2024-04-15T11-01-11.453575.parquet
- 分割: latest
- 路径: **/details_harness|hellaswag|10_2024-04-15T11-01-11.453575.parquet
- 分割: 2024_04_15T11_01_11.453575
-
config_name: harness_hendrycksTest_5
- 分割: 2024_04_15T11_01_11.453575
- 路径:
- **/details_harness|hendrycksTest-abstract_algebra|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-anatomy|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-astronomy|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-business_ethics|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-clinical_knowledge|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-college_biology|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-college_chemistry|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-college_computer_science|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-college_mathematics|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-college_medicine|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-college_physics|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-computer_security|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-conceptual_physics|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-econometrics|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-electrical_engineering|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-elementary_mathematics|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-formal_logic|5_2024-04-15T11-01-11.453575.parquet
- **/details_harness|hendrycksTest-global_facts|5_2024-04-15T11-01-11.453575.parquet
其他路径...
- 路径:
- 分割: 2024_04_15T11_01_11.453575
}



