open-llm-leaderboard/details_OpenBuddy__openbuddy-codellama2-34b-v11.1-bf16
收藏数据集概述
数据集简介
该数据集是在评估模型 OpenBuddy/openbuddy-codellama2-34b-v11.1-bf16 在 Open LLM Leaderboard 上的自动创建的。数据集包含64个配置,每个配置对应一个评估任务。
数据集结构
数据集由3次运行结果组成,每次运行的结果可以在每个配置中找到,并以运行的时间戳命名。每个配置中的 "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
以下是加载特定运行结果的示例代码: python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_OpenBuddy__openbuddy-codellama2-34b-v11.1-bf16", "harness_winogrande_5", split="train")
最新结果
以下是 2023-10-28T03:42:28.997128 运行 的最新结果: python { "all": { "em": 0.360633389261745, "em_stderr": 0.004917536525106699, "f1": 0.4180935402684579, "f1_stderr": 0.004778710905980245, "acc": 0.5268440191410464, "acc_stderr": 0.012939810741097795 }, "harness|drop|3": { "em": 0.360633389261745, "em_stderr": 0.004917536525106699, "f1": 0.4180935402684579, "f1_stderr": 0.004778710905980245 }, "harness|gsm8k|5": { "acc": 0.3457164518574678, "acc_stderr": 0.013100422990441578 }, "harness|winogrande|5": { "acc": 0.7079715864246251, "acc_stderr": 0.012779198491754013 } }
配置详情
以下是数据集的配置详情:
-
harness_arc_challenge_25
- 分割: 2023_10_03T23_40_22.620996
- 路径:
**/details_harness|arc:challenge|25_2023-10-03T23-40-22.620996.parquet
- 路径:
- 分割: latest
- 路径:
**/details_harness|arc:challenge|25_2023-10-03T23-40-22.620996.parquet
- 路径:
- 分割: 2023_10_03T23_40_22.620996
-
harness_drop_3
- 分割: 2023_10_27T21_47_43.594265
- 路径:
**/details_harness|drop|3_2023-10-27T21-47-43.594265.parquet
- 路径:
- 分割: 2023_10_28T03_42_28.997128
- 路径:
**/details_harness|drop|3_2023-10-28T03-42-28.997128.parquet
- 路径:
- 分割: latest
- 路径:
**/details_harness|drop|3_2023-10-28T03-42-28.997128.parquet
- 路径:
- 分割: 2023_10_27T21_47_43.594265
-
harness_gsm8k_5
- 分割: 2023_10_27T21_47_43.594265
- 路径:
**/details_harness|gsm8k|5_2023-10-27T21-47-43.594265.parquet
- 路径:
- 分割: 2023_10_28T03_42_28.997128
- 路径:
**/details_harness|gsm8k|5_2023-10-28T03-42-28.997128.parquet
- 路径:
- 分割: latest
- 路径:
**/details_harness|gsm8k|5_2023-10-28T03-42-28.997128.parquet
- 路径:
- 分割: 2023_10_27T21_47_43.594265
-
harness_hellaswag_10
- 分割: 2023_10_03T23_40_22.620996
- 路径:
**/details_harness|hellaswag|10_2023-10-03T23-40-22.620996.parquet
- 路径:
- 分割: latest
- 路径:
**/details_harness|hellaswag|10_2023-10-03T23-40-22.620996.parquet
- 路径:
- 分割: 2023_10_03T23_40_22.620996
-
harness_hendrycksTest_5
- 分割: 2023_10_03T23_40_22.620996
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-formal_logic|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-global_facts|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-high_school_biology|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-high_school_european_history|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-high_school_geography|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-high_school_physics|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-high_school_psychology|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-high_school_statistics|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-high_school_us_history|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-high_school_world_history|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-human_aging|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-human_sexuality|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-international_law|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-jurisprudence|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-logical_fallacies|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-machine_learning|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-management|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-marketing|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-medical_genetics|5_2023-10-03T23-40-22.620996.parquet - 路径:
**/details_harness|hendrycksTest-miscellaneous|5_2023-10-03T23-40-22.620996.parquet - 路径: `**/details_harness|hendrycksTest-moral_disputes|5_2
- 路径:
- 分割: 2023_10_03T23_40_22.620996



