open-llm-leaderboard-old/details_tiiuae__falcon-180B
收藏数据集概述
数据集摘要
该数据集是在对模型 tiiuae/falcon-180B 进行评估运行期间自动创建的,用于 Open LLM Leaderboard。
数据集组成
- 数据集包含 66 个配置,每个配置对应一个评估任务。
- 数据集由 32 次运行创建,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。
- 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_tiiuae__falcon-180B", "harness_winogrande_5", split="train")
最新结果
以下是 2023-10-24T10:17:51.759984 运行的最新结果: python { "all": { "em": 0.0028313758389261743, "em_stderr": 0.0005441551135493806, "f1": 0.06573301174496615, "f1_stderr": 0.0013666874377791776, "acc": 0.6642104078991223, "acc_stderr": 0.011605139145295384 }, "harness|drop|3": { "em": 0.0028313758389261743, "em_stderr": 0.0005441551135493806, "f1": 0.06573301174496615, "f1_stderr": 0.0013666874377791776 }, "harness|gsm8k|5": { "acc": 0.45943896891584535, "acc_stderr": 0.01372709301042978 }, "harness|winogrande|5": { "acc": 0.8689818468823993, "acc_stderr": 0.009483185280160986 } }
配置详情
-
config_name: harness_arc_challenge_25
- 数据文件:
- split: 2023_08_30T14_31_39.488381
- path: **/details_harness|arc:challenge|25_2023-08-30T14:31:39.488381.parquet
- split: 2023_08_30T19_27_57.090829
- path: **/details_harness|arc:challenge|25_2023-08-30T19:27:57.090829.parquet
- split: 2023_08_31T01_32_36.577851
- path: **/details_harness|arc:challenge|25_2023-08-31T01:32:36.577851.parquet
- split: 2023_08_31T12_44_38.148712
- path: **/details_harness|arc:challenge|25_2023-08-31T12:44:38.148712.parquet
- split: 2023_09_01T15_12_02.263774
- path: **/details_harness|arc:challenge|25_2023-09-01T15:12:02.263774.parquet
- split: 2023_09_25T09_30_46.601936
- path: **/details_harness|arc:challenge|25_2023-09-25T09-30-46.601936.parquet
- split: 2023_09_25T09_42_43.006060
- path: **/details_harness|arc:challenge|25_2023-09-25T09-42-43.006060.parquet
- split: latest
- path: **/details_harness|arc:challenge|25_2023-09-25T09-42-43.006060.parquet
- split: 2023_08_30T14_31_39.488381
- 数据文件:
-
config_name: harness_drop_3
- 数据文件:
- split: 2023_10_23T17_29_05.444286
- path: **/details_harness|drop|3_2023-10-23T17-29-05.444286.parquet
- split: 2023_10_24T10_17_51.759984
- path: **/details_harness|drop|3_2023-10-24T10-17-51.759984.parquet
- split: latest
- path: **/details_harness|drop|3_2023-10-24T10-17-51.759984.parquet
- split: 2023_10_23T17_29_05.444286
- 数据文件:
-
config_name: harness_gsm8k_5
- 数据文件:
- split: 2023_10_23T17_29_05.444286
- path: **/details_harness|gsm8k|5_2023-10-23T17-29-05.444286.parquet
- split: 2023_10_24T10_17_51.759984
- path: **/details_harness|gsm8k|5_2023-10-24T10-17-51.759984.parquet
- split: latest
- path: **/details_harness|gsm8k|5_2023-10-24T10-17-51.759984.parquet
- split: 2023_10_23T17_29_05.444286
- 数据文件:
-
config_name: harness_hellaswag_10
- 数据文件:
- split: 2023_08_30T14_31_39.488381
- path: **/details_harness|hellaswag|10_2023-08-30T14:31:39.488381.parquet
- split: 2023_08_30T19_27_57.090829
- path: **/details_harness|hellaswag|10_2023-08-30T19:27:57.090829.parquet
- split: 2023_08_31T01_32_36.577851
- path: **/details_harness|hellaswag|10_2023-08-31T01:32:36.577851.parquet
- split: 2023_08_31T12_44_38.148712
- path: **/details_harness|hellaswag|10_2023-08-31T12:44:38.148712.parquet
- split: 2023_09_01T15_12_02.263774
- path: **/details_harness|hellaswag|10_2023-09-01T15:12:02.263774.parquet
- split: 2023_09_25T11_16_10.146827
- path: **/details_harness|hellaswag|10_2023-09-25T11-16-10.146827.parquet
- split: 2023_09_25T11_28_53.879118
- path: **/details_harness|hellaswag|10_2023-09-25T11-28-53.879118.parquet
- split: 2023_09_25T13_20_00.898508
- path: **/details_harness|hellaswag|10_2023-09-25T13-20-00.898508.parquet
- split: latest
- path: **/details_harness|hellaswag|10_2023-09-25T13-20-00.898508.parquet
- split: 2023_08_30T14_31_39.488381
- 数据文件:
-
config_name: harness_hendrycksTest_5
- 数据文件:
- split: 2023_08_30T14_31_39.488381
- path: **/details_harness|hendrycksTest-abstract_algebra|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-anatomy|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-astronomy|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-business_ethics|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-college_biology|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-college_chemistry|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-college_computer_science|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-college_mathematics|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-college_medicine|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-college_physics|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-computer_security|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-conceptual_physics|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-econometrics|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-electrical_engineering|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-formal_logic|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-global_facts|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-high_school_biology|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-high_school_european_history|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-high_school_geography|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-30T14:31:39.488381.parquet
- path: **/details_harness|hendrycksTest-high_school_macroeconomics|5_20
- split: 2023_08_30T14_31_39.488381
- 数据文件:



