open-llm-leaderboard-old/details_YeungNLP__firefly-llama-30b
收藏数据集概述
数据集简介
该数据集是在评估模型 YeungNLP/firefly-llama-30b 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
- 配置数量: 64
- 数据来源: 2 次运行结果
- 数据格式: Parquet 文件
- 数据分割: 每个配置包含特定运行的分割,分割名称使用运行的时间戳。"train" 分割指向最新的结果。
特殊配置
- 结果配置: 名为 "results" 的配置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_YeungNLP__firefly-llama-30b", "harness_winogrande_5", split="train")
最新结果
以下是 2023-10-16T19:25:26.382420 运行的最新结果: python { "all": { "em": 0.3185822147651007, "em_stderr": 0.004771524379931377, "f1": 0.39832948825503617, "f1_stderr": 0.004636207519012633, "acc": 0.46636165136703384, "acc_stderr": 0.010904050525570757 }, "harness|drop|3": { "em": 0.3185822147651007, "em_stderr": 0.004771524379931377, "f1": 0.39832948825503617, "f1_stderr": 0.004636207519012633 }, "harness|gsm8k|5": { "acc": 0.15845337376800606, "acc_stderr": 0.010058474790238966 }, "harness|winogrande|5": { "acc": 0.7742699289660616, "acc_stderr": 0.011749626260902549 } }
配置详情
以下是部分配置的详细信息:
-
harness_arc_challenge_25:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|arc:challenge|25_2023-08-10T11:37:52.029669.parquet
-
harness_drop_3:
- 分割: 2023_10_16T19_25_26.382420, latest
- 路径:
**/details_harness|drop|3_2023-10-16T19-25-26.382420.parquet
-
harness_gsm8k_5:
- 分割: 2023_10_16T19_25_26.382420, latest
- 路径:
**/details_harness|gsm8k|5_2023-10-16T19-25-26.382420.parquet
-
harness_hellaswag_10:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hellaswag|10_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径: 多个路径,包括
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-10T11:37:52.029669.parquet等
-
harness_hendrycksTest_abstract_algebra_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_anatomy_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_astronomy_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_business_ethics_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_clinical_knowledge_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_college_biology_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_college_chemistry_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_college_computer_science_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_college_mathematics_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_college_medicine_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_college_physics_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_computer_security_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_conceptual_physics_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_econometrics_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_electrical_engineering_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-10T11:37:52.029669.parquet
-
harness_hendrycksTest_elementary_mathematics_5:
- 分割: 2023_08_10T11_37_52.029669, latest
- 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-10T11:37:52.029669.parquet
以上是数据集的详细概述,包括数据集的简介、结构、特殊配置、数据加载示例、最新结果以及部分配置的详细信息。



