open-llm-leaderboard/details_facebook__opt-350m
收藏数据集概述
该数据集是在评估模型 facebook/opt-350m 在 Open LLM Leaderboard 上的运行过程中自动创建的。
数据集组成
- 数据集包含 64 个配置,每个配置对应一个评估任务。
- 数据集从 3 次运行中创建,每次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。
- 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 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_facebook__opt-350m", "harness_winogrande_5", split="train")
最新结果
以下是 2023-10-29T05:40:05.173534 运行的最新结果: python { "all": { "em": 0.0006291946308724832, "em_stderr": 0.0002568002749723937, "f1": 0.04159815436241622, "f1_stderr": 0.0011509154641292957, "acc": 0.26473650543660054, "acc_stderr": 0.007773698717815887 }, "harness|drop|3": { "em": 0.0006291946308724832, "em_stderr": 0.0002568002749723937, "f1": 0.04159815436241622, "f1_stderr": 0.0011509154641292957 }, "harness|gsm8k|5": { "acc": 0.003032600454890068, "acc_stderr": 0.0015145735612245468 }, "harness|winogrande|5": { "acc": 0.526440410418311, "acc_stderr": 0.014032823874407229 } }
配置详情
以下是数据集的配置详情:
-
config_name: harness_arc_challenge_25
- split: 2023_07_24T09_55_50.700214
- path: **/details_harness|arc:challenge|25_2023-07-24T09:55:50.700214.parquet
- split: latest
- path: **/details_harness|arc:challenge|25_2023-07-24T09:55:50.700214.parquet
- split: 2023_07_24T09_55_50.700214
-
config_name: harness_drop_3
- split: 2023_10_24T09_18_34.693204
- path: **/details_harness|drop|3_2023-10-24T09-18-34.693204.parquet
- split: 2023_10_29T05_40_05.173534
- path: **/details_harness|drop|3_2023-10-29T05-40-05.173534.parquet
- split: latest
- path: **/details_harness|drop|3_2023-10-29T05-40-05.173534.parquet
- split: 2023_10_24T09_18_34.693204
-
config_name: harness_gsm8k_5
- split: 2023_10_24T09_18_34.693204
- path: **/details_harness|gsm8k|5_2023-10-24T09-18-34.693204.parquet
- split: 2023_10_29T05_40_05.173534
- path: **/details_harness|gsm8k|5_2023-10-29T05-40-05.173534.parquet
- split: latest
- path: **/details_harness|gsm8k|5_2023-10-29T05-40-05.173534.parquet
- split: 2023_10_24T09_18_34.693204
-
config_name: harness_hellaswag_10
- split: 2023_07_24T09_55_50.700214
- path: **/details_harness|hellaswag|10_2023-07-24T09:55:50.700214.parquet
- split: latest
- path: **/details_harness|hellaswag|10_2023-07-24T09:55:50.700214.parquet
- split: 2023_07_24T09_55_50.700214
-
config_name: harness_hendrycksTest_5
- split: 2023_07_24T09_55_50.700214
- path:
- **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-anatomy|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-astronomy|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-business_ethics|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-college_biology|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-college_chemistry|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-college_computer_science|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-college_mathematics|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-college_medicine|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-college_physics|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-computer_security|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-conceptual_physics|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-econometrics|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-electrical_engineering|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-formal_logic|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-global_facts|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-high_school_biology|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-high_school_european_history|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-high_school_geography|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-high_school_physics|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-high_school_psychology|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-high_school_statistics|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-high_school_us_history|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-high_school_world_history|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-human_aging|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-human_sexuality|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-international_law|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-jurisprudence|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-logical_fallacies|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-machine_learning|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-management|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-marketing|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-medical_genetics|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-miscellaneous|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-moral_disputes|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-moral_scenarios|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-nutrition|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-philosophy|5_2023-07-24T09:55:50.700214.parquet
- **/details_harness|hendrycksTest-prehistory|
- path:
- split: 2023_07_24T09_55_50.700214




