open-llm-leaderboard-old/details_eachadea__vicuna-7b-1.1
收藏数据集概述
数据集简介
该数据集是在评估模型 eachadea/vicuna-7b-1.1 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:数据集从2次运行中创建。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" 分割总是指向最新的结果。
- 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_eachadea__vicuna-7b-1.1", "harness_winogrande_5", split="train")
最新结果
以下是 2023-09-22T23:37:12.210643 运行 的最新结果: python { "all": { "em": 0.11388422818791946, "em_stderr": 0.00325324428862373, "f1": 0.16976719798657605, "f1_stderr": 0.003380156230610554, "acc": 0.38244753834582057, "acc_stderr": 0.009528517622122097 }, "harness|drop|3": { "em": 0.11388422818791946, "em_stderr": 0.00325324428862373, "f1": 0.16976719798657605, "f1_stderr": 0.003380156230610554 }, "harness|gsm8k|5": { "acc": 0.05534495830174375, "acc_stderr": 0.006298221796179588 }, "harness|winogrande|5": { "acc": 0.7095501183898973, "acc_stderr": 0.012758813448064607 } }
配置详情
以下是部分配置及其数据文件路径:
-
harness_arc_challenge_25
- 分割:2023_07_18T12_22_46.451039
- 路径:
**/details_harness|arc:challenge|25_2023-07-18T12:22:46.451039.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|arc:challenge|25_2023-07-18T12:22:46.451039.parquet
- 路径:
- 分割:2023_07_18T12_22_46.451039
-
harness_drop_3
- 分割:2023_09_22T23_37_12.210643
- 路径:
**/details_harness|drop|3_2023-09-22T23-37-12.210643.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|drop|3_2023-09-22T23-37-12.210643.parquet
- 路径:
- 分割:2023_09_22T23_37_12.210643
-
harness_gsm8k_5
- 分割:2023_09_22T23_37_12.210643
- 路径:
**/details_harness|gsm8k|5_2023-09-22T23-37-12.210643.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|gsm8k|5_2023-09-22T23-37-12.210643.parquet
- 路径:
- 分割:2023_09_22T23_37_12.210643
-
harness_hellaswag_10
- 分割:2023_07_18T12_22_46.451039
- 路径:
**/details_harness|hellaswag|10_2023-07-18T12:22:46.451039.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|hellaswag|10_2023-07-18T12:22:46.451039.parquet
- 路径:
- 分割:2023_07_18T12_22_46.451039
-
harness_hendrycksTest_5
- 分割:2023_07_18T12_22_46.451039
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-formal_logic|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-global_facts|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-high_school_biology|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-high_school_european_history|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-high_school_geography|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-high_school_physics|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-high_school_psychology|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-high_school_statistics|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-high_school_us_history|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-high_school_world_history|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-human_aging|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-human_sexuality|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-international_law|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-jurisprudence|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-logical_fallacies|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-machine_learning|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-management|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-marketing|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-medical_genetics|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-miscellaneous|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-moral_disputes|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-moral_scenarios|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-nutrition|5_2023-07-18T12:22:46.451039.parquet - 路径:
**/details_harness|hendrycksTest-philosophy|5_2023-07-18T12:22:46.451039.parquet - 路径:`**/details_harness|hendrycksTest-
- 路径:
- 分割:2023_07_18T12_22_46.451039



