open-llm-leaderboard/details_TinyPixel__llama2-7b-instruct
收藏数据集概述
数据集简介
该数据集是在评估模型 TinyPixel/llama2-7b-instruct 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 61 个配置,每个配置对应一个评估任务。
- 数据集从 1 次运行中创建,每个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 "results" 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_TinyPixel__llama2-7b-instruct", "harness_truthfulqa_mc_0", split="train")
最新结果
以下是 2023-08-17T12:12:37.965756 运行的最新结果:
python { "all": { "acc": 0.4639503533482998, "acc_stderr": 0.03519400615590806, "acc_norm": 0.467921814589003, "acc_norm_stderr": 0.03517936985393269, "mc1": 0.26438188494492043, "mc1_stderr": 0.015438211119522512, "mc2": 0.39481096196846566, "mc2_stderr": 0.013796205321597201 }, "harness|arc:challenge|25": { "acc": 0.49829351535836175, "acc_stderr": 0.01461130570505699, "acc_norm": 0.5358361774744027, "acc_norm_stderr": 0.01457381366473572 }, "harness|hellaswag|10": { "acc": 0.5910177255526787, "acc_stderr": 0.004906411984476793, "acc_norm": 0.7877912766381199, "acc_norm_stderr": 0.00408036220825117 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.3, "acc_stderr": 0.046056618647183814, "acc_norm": 0.3, "acc_norm_stderr": 0.046056618647183814 }, # 其他任务的结果省略... }
配置详情
-
config_name: harness_arc_challenge_25
- 数据文件:
- 分割: 2023_08_17T12_12_37.965756
- 路径: **/details_harness|arc:challenge|25_2023-08-17T12:12:37.965756.parquet
- 分割: latest
- 路径: **/details_harness|arc:challenge|25_2023-08-17T12:12:37.965756.parquet
- 分割: 2023_08_17T12_12_37.965756
- 数据文件:
-
config_name: harness_hellaswag_10
- 数据文件:
- 分割: 2023_08_17T12_12_37.965756
- 路径: **/details_harness|hellaswag|10_2023-08-17T12:12:37.965756.parquet
- 分割: latest
- 路径: **/details_harness|hellaswag|10_2023-08-17T12:12:37.965756.parquet
- 分割: 2023_08_17T12_12_37.965756
- 数据文件:
-
config_name: harness_hendrycksTest_5
- 数据文件:
- 分割: 2023_08_17T12_12_37.965756
- 路径:
- **/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-anatomy|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-astronomy|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-business_ethics|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-college_biology|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-college_chemistry|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-college_computer_science|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-college_mathematics|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-college_medicine|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-college_physics|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-computer_security|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-conceptual_physics|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-econometrics|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-electrical_engineering|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-formal_logic|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-global_facts|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-high_school_biology|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-high_school_european_history|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-high_school_geography|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-high_school_physics|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-high_school_psychology|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-high_school_statistics|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-high_school_us_history|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-high_school_world_history|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-human_aging|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-human_sexuality|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-international_law|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-jurisprudence|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-logical_fallacies|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-machine_learning|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-management|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-marketing|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-medical_genetics|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-miscellaneous|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-moral_disputes|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-moral_scenarios|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-nutrition|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-philosophy|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-prehistory|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-professional_accounting|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-professional_law|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-professional_medicine|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-professional_psychology|5_2023-08-17T12:12:37.965756.parquet
- **/details_harness|hendrycksTest-public_relations|5_2023
- 路径:
- 分割: 2023_08_17T12_12_37.965756
- 数据文件:



