open-llm-leaderboard-old/details_PulsarAI__EnsembleV5-Nova-13B

该数据集是在模型PulsarAI/EnsembleV5-Nova-13B在Open LLM Leaderboard上的评估运行过程中自动创建的。数据集由64个配置组成，每个配置对应一个评估任务。数据集由2次运行生成，每次运行的结果作为特定配置中的一个split，split以运行的时间戳命名。train split始终指向最新的结果。此外，还有一个名为results的配置存储了所有运行的聚合结果，用于在Open LLM Leaderboard上计算和显示聚合指标。

原始信息汇总

数据集概述

数据集来源

该数据集是在模型 PulsarAI/EnsembleV5-Nova-13B 在 Open LLM Leaderboard 上的评估运行期间自动创建的。

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 2 次运行中创建，每次运行可以在每个配置中作为一个特定的分片找到，分片名称使用运行的时间戳。
• "train" 分片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_PulsarAI__EnsembleV5-Nova-13B", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-23T15:24:00.966689 运行的最新结果： python { "all": { "em": 0.007445469798657718, "em_stderr": 0.0008803652515899855, "f1": 0.08636220637583875, "f1_stderr": 0.0018310737230495444, "acc": 0.4350441276875584, "acc_stderr": 0.010249391454413254 }, "harness|drop|3": { "em": 0.007445469798657718, "em_stderr": 0.0008803652515899855, "f1": 0.08636220637583875, "f1_stderr": 0.0018310737230495444 }, "harness|gsm8k|5": { "acc": 0.10765731614859743, "acc_stderr": 0.008537484003023352 }, "harness|winogrande|5": { "acc": 0.7624309392265194, "acc_stderr": 0.011961298905803157 } }

配置详情

以下是部分配置的详细信息：

• config_name: harness_arc_challenge_25

  • split: 2023_10_03T19_22_59.151966
    • path: **/details_harness|arc:challenge|25_2023-10-03T19-22-59.151966.parquet
  • split: latest
    • path: **/details_harness|arc:challenge|25_2023-10-03T19-22-59.151966.parquet

• config_name: harness_drop_3

  • split: 2023_10_23T15_24_00.966689
    • path: **/details_harness|drop|3_2023-10-23T15-24-00.966689.parquet
  • split: latest
    • path: **/details_harness|drop|3_2023-10-23T15-24-00.966689.parquet

• config_name: harness_gsm8k_5

  • split: 2023_10_23T15_24_00.966689
    • path: **/details_harness|gsm8k|5_2023-10-23T15-24-00.966689.parquet
  • split: latest
    • path: **/details_harness|gsm8k|5_2023-10-23T15-24-00.966689.parquet

• config_name: harness_hellaswag_10

  • split: 2023_10_03T19_22_59.151966
    • path: **/details_harness|hellaswag|10_2023-10-03T19-22-59.151966.parquet
  • split: latest
    • path: **/details_harness|hellaswag|10_2023-10-03T19-22-59.151966.parquet

• config_name: harness_hendrycksTest_5

  • split: 2023_10_03T19_22_59.151966
    • path:
      • **/details_harness|hendrycksTest-abstract_algebra|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-anatomy|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-astronomy|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-business_ethics|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-college_biology|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-college_chemistry|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-college_computer_science|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-college_mathematics|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-college_medicine|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-college_physics|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-computer_security|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-conceptual_physics|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-econometrics|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-electrical_engineering|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-formal_logic|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-global_facts|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-high_school_biology|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-high_school_european_history|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-high_school_geography|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-high_school_physics|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-high_school_psychology|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-high_school_statistics|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-high_school_us_history|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-high_school_world_history|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-human_aging|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-human_sexuality|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-international_law|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-jurisprudence|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-logical_fallacies|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-machine_learning|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-management|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-marketing|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-medical_genetics|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-miscellaneous|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-moral_disputes|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-moral_scenarios|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-nutrition|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-philosophy|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-prehistory|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-professional_accounting|5_2023-10-03T19-22-59.151966.parquet
      • **/details_harness|hendrycksTest-professional_law|5_2023-10-03T19-22-59.