open-llm-leaderboard/details_OpenBuddy__openbuddy-atom-13b-v9-bf16
收藏数据集概述
数据集简介
该数据集是在对模型 OpenBuddy/openbuddy-atom-13b-v9-bf16 进行评估运行期间自动创建的,用于 Open LLM Leaderboard。
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。
- 数据来源:数据集由2次运行创建,每次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" 分割始终指向最新结果。
- 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 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_OpenBuddy__openbuddy-atom-13b-v9-bf16", "harness_winogrande_5", split="train")
最新结果
以下是 2023-10-15T21:37:39.062296 运行 的最新结果: python { "all": { "em": 0.057466442953020135, "em_stderr": 0.0023833905882384974, "f1": 0.11402369966442945, "f1_stderr": 0.0026622077831256583, "acc": 0.44356628547732635, "acc_stderr": 0.011184922703096678 }, "harness|drop|3": { "em": 0.057466442953020135, "em_stderr": 0.0023833905882384974, "f1": 0.11402369966442945, "f1_stderr": 0.0026622077831256583 }, "harness|gsm8k|5": { "acc": 0.15390447308567096, "acc_stderr": 0.00993979930404902 }, "harness|winogrande|5": { "acc": 0.7332280978689818, "acc_stderr": 0.012430046102144337 } }
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_08_17T18_31_32.257089
- 路径:
**/details_harness|arc:challenge|25_2023-08-17T18:31:32.257089.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|arc:challenge|25_2023-08-17T18:31:32.257089.parquet
- 路径:
- 分割:2023_08_17T18_31_32.257089
- harness_drop_3
- 分割:2023_10_15T21_37_39.062296
- 路径:
**/details_harness|drop|3_2023-10-15T21-37-39.062296.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|drop|3_2023-10-15T21-37-39.062296.parquet
- 路径:
- 分割:2023_10_15T21_37_39.062296
- harness_gsm8k_5
- 分割:2023_10_15T21_37_39.062296
- 路径:
**/details_harness|gsm8k|5_2023-10-15T21-37-39.062296.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|gsm8k|5_2023-10-15T21-37-39.062296.parquet
- 路径:
- 分割:2023_10_15T21_37_39.062296
- harness_hellaswag_10
- 分割:2023_08_17T18_31_32.257089
- 路径:
**/details_harness|hellaswag|10_2023-08-17T18:31:32.257089.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|hellaswag|10_2023-08-17T18:31:32.257089.parquet
- 路径:
- 分割:2023_08_17T18_31_32.257089
- harness_hendrycksTest_5
- 分割:2023_08_17T18_31_32.257089
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-formal_logic|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-global_facts|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-high_school_biology|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-high_school_geography|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-high_school_physics|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-high_school_psychology|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-high_school_statistics|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-high_school_us_history|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-high_school_world_history|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-human_aging|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-human_sexuality|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-international_law|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-jurisprudence|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-logical_fallacies|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-machine_learning|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-management|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-marketing|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-medical_genetics|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-miscellaneous|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-moral_disputes|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-moral_scenarios|5_2023-08-17T18:31:32.257089.parquet - 路径:
**/details_harness|hendrycksTest-nutrition|5_2023-08-17T18:31:32.257089.parquet - 路径:`**/details_harness|hendrycksTest-philosophy|5_2023-08-17T18:31:32.
- 路径:
- 分割:2023_08_17T18_31_32.257089



