open-llm-leaderboard-old/details_lgaalves__gpt2-dolly
收藏数据集概述
数据集来源
该数据集是在对模型 lgaalves/gpt2-dolly 进行评估运行期间自动创建的,评估运行在 Open LLM Leaderboard 上进行。
数据集结构
- 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 4 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_lgaalves__gpt2-dolly", "harness_winogrande_5", split="train")
最新结果
以下是 2023-10-26T15:16:18.909977 运行的最新结果: python { "all": { "em": 0.0008389261744966443, "em_stderr": 0.00029649629898012396, "f1": 0.034500838926174546, "f1_stderr": 0.0010901499685640162, "acc": 0.25805886045310694, "acc_stderr": 0.007559135865912546 }, "harness|drop|3": { "em": 0.0008389261744966443, "em_stderr": 0.00029649629898012396, "f1": 0.034500838926174546, "f1_stderr": 0.0010901499685640162 }, "harness|gsm8k|5": { "acc": 0.001516300227445034, "acc_stderr": 0.0010717793485492627 }, "harness|winogrande|5": { "acc": 0.5146014206787688, "acc_stderr": 0.01404649238327583 } }
配置详情
-
harness_arc_challenge_25
- 分割:2023_08_09T12_04_01.298115
- 路径:
**/details_harness|arc:challenge|25_2023-08-09T12:04:01.298115.parquet
- 路径:
- 分割:2023_09_21T18_57_43.248355
- 路径:
**/details_harness|arc:challenge|25_2023-09-21T18-57-43.248355.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|arc:challenge|25_2023-09-21T18-57-43.248355.parquet
- 路径:
- 分割:2023_08_09T12_04_01.298115
-
harness_drop_3
- 分割:2023_10_16T15_21_03.106621
- 路径:
**/details_harness|drop|3_2023-10-16T15-21-03.106621.parquet
- 路径:
- 分割:2023_10_26T15_16_18.909977
- 路径:
**/details_harness|drop|3_2023-10-26T15-16-18.909977.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|drop|3_2023-10-26T15-16-18.909977.parquet
- 路径:
- 分割:2023_10_16T15_21_03.106621
-
harness_gsm8k_5
- 分割:2023_10_16T15_21_03.106621
- 路径:
**/details_harness|gsm8k|5_2023-10-16T15-21-03.106621.parquet
- 路径:
- 分割:2023_10_26T15_16_18.909977
- 路径:
**/details_harness|gsm8k|5_2023-10-26T15-16-18.909977.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|gsm8k|5_2023-10-26T15-16-18.909977.parquet
- 路径:
- 分割:2023_10_16T15_21_03.106621
-
harness_hellaswag_10
- 分割:2023_08_09T12_04_01.298115
- 路径:
**/details_harness|hellaswag|10_2023-08-09T12:04:01.298115.parquet
- 路径:
- 分割:2023_09_21T18_57_43.248355
- 路径:
**/details_harness|hellaswag|10_2023-09-21T18-57-43.248355.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|hellaswag|10_2023-09-21T18-57-43.248355.parquet
- 路径:
- 分割:2023_08_09T12_04_01.298115
-
harness_hendrycksTest_5
- 分割:2023_08_09T12_04_01.298115
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-formal_logic|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-global_facts|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-high_school_biology|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-high_school_geography|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-high_school_physics|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-high_school_psychology|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-high_school_statistics|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-high_school_us_history|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-high_school_world_history|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-human_aging|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-human_sexuality|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-international_law|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-jurisprudence|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-logical_fallacies|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-machine_learning|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-management|5_2023-08-09T12:04:01.298115.parquet - 路径:
**/details_harness|hendrycksTest-marketing|5_2023-08-09T12:04:01.298115.parquet - 路径:`**/details_harness|hendrycksTest-medical_genetics|5_2023-08-
- 路径:
- 分割:2023_08_09T12_04_01.298115



