open-llm-leaderboard-old/details_stabilityai__stablelm-tuned-alpha-7b
收藏数据集概述
数据集名为 "Evaluation run of stabilityai/stablelm-tuned-alpha-7b",是在评估模型 stabilityai/stablelm-tuned-alpha-7b 时自动创建的。该数据集用于 Open LLM Leaderboard 的评估。
数据集组成
数据集由 64 个配置组成,每个配置对应一个评估任务。数据集从 2 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。
额外配置
一个额外的配置 "results" 存储了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
以下是加载特定运行详细信息的示例代码:
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_stabilityai__stablelm-tuned-alpha-7b", "harness_winogrande_5", split="train")
最新结果
以下是 最新结果 的示例:
python { "all": { "em": 0.0041946308724832215, "em_stderr": 0.0006618716168266466, "f1": 0.05621224832214779, "f1_stderr": 0.0014117433231649174, "acc": 0.2697578287825378, "acc_stderr": 0.008265042433750026 }, "harness|drop|3": { "em": 0.0041946308724832215, "em_stderr": 0.0006618716168266466, "f1": 0.05621224832214779, "f1_stderr": 0.0014117433231649174 }, "harness|gsm8k|5": { "acc": 0.008339651250947688, "acc_stderr": 0.002504942226860537 }, "harness|winogrande|5": { "acc": 0.5311760063141279, "acc_stderr": 0.014025142640639513 } }
配置详情
数据集包含多个配置,每个配置对应不同的任务和数据文件路径。以下是部分配置的示例:
-
harness_arc_challenge_25
- 分割:2023_07_19T17_04_40.596532
- 路径:
**/details_harness|arc:challenge|25_2023-07-19T17:04:40.596532.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|arc:challenge|25_2023-07-19T17:04:40.596532.parquet
- 路径:
- 分割:2023_07_19T17_04_40.596532
-
harness_drop_3
- 分割:2023_10_12T19_40_34.606567
- 路径:
**/details_harness|drop|3_2023-10-12T19-40-34.606567.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|drop|3_2023-10-12T19-40-34.606567.parquet
- 路径:
- 分割:2023_10_12T19_40_34.606567
-
harness_gsm8k_5
- 分割:2023_10_12T19_40_34.606567
- 路径:
**/details_harness|gsm8k|5_2023-10-12T19-40-34.606567.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|gsm8k|5_2023-10-12T19-40-34.606567.parquet
- 路径:
- 分割:2023_10_12T19_40_34.606567
-
harness_hellaswag_10
- 分割:2023_07_19T17_04_40.596532
- 路径:
**/details_harness|hellaswag|10_2023-07-19T17:04:40.596532.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|hellaswag|10_2023-07-19T17:04:40.596532.parquet
- 路径:
- 分割:2023_07_19T17_04_40.596532
-
harness_hendrycksTest_5
- 分割:2023_07_19T17_04_40.596532
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-formal_logic|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-global_facts|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-high_school_physics|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-high_school_psychology|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-high_school_statistics|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-high_school_us_history|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-high_school_world_history|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-human_aging|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-human_sexuality|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-international_law|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-jurisprudence|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-logical_fallacies|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-machine_learning|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-management|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-marketing|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-medical_genetics|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-miscellaneous|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-moral_disputes|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-moral_scenarios|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-nutrition|5_2023-07-19T17:04:40.596532.parquet - 路径:
**/details_harness|hendrycksTest-philosophy|5_2023-07-19T17:04:40.596532.parquet
- 路径:
- 分割:2023_07_19T17_04_40.596532



