five

open-llm-leaderboard/details_LLMs__WizardLM-13B-V1.0

收藏
Hugging Face2023-09-23 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_LLMs__WizardLM-13B-V1.0
下载链接
链接失效反馈
官方服务:
资源简介:
该数据集是在Open LLM Leaderboard上对模型LLMs/WizardLM-13B-V1.0进行评估时自动创建的。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。此外,还有一个名为results的配置,存储了所有运行的聚合结果,并用于计算和显示Open LLM Leaderboard上的聚合指标。

This dataset was automatically generated during the evaluation of the LLMs/WizardLM-13B-V1.0 model on the Open LLM Leaderboard. It comprises 64 configurations, each corresponding to an individual evaluation task. The dataset is produced via two experimental runs, where the results of each run act as a split under its respective configuration, with the split name being the timestamp of the corresponding run. The `train` split always references the most up-to-date results. Additionally, there exists a configuration titled `results` that stores the aggregated results across all runs, and is utilized to compute and display the aggregated metrics featured on the Open LLM Leaderboard.
提供机构:
open-llm-leaderboard
原始信息汇总

数据集概述

数据集名称

  • Evaluation run of LLMs/WizardLM-13B-V1.0

数据集描述

数据集组成

  • 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_LLMs__WizardLM-13B-V1.0", "harness_winogrande_5", split="train")

最新结果

  • 这些是最新结果(来自 2023-09-23T04:14:08.122970 运行): python { "all": { "em": 0.06669463087248322, "em_stderr": 0.0025550356847442663, "f1": 0.1407225251677848, "f1_stderr": 0.002886547618779489, "acc": 0.44106833942386575, "acc_stderr": 0.010948605580118738 }, "harness|drop|3": { "em": 0.06669463087248322, "em_stderr": 0.0025550356847442663, "f1": 0.1407225251677848, "f1_stderr": 0.002886547618779489 }, "harness|gsm8k|5": { "acc": 0.14101592115238817, "acc_stderr": 0.009586695349244096 }, "harness|winogrande|5": { "acc": 0.7411207576953434, "acc_stderr": 0.01231051581099338 } }

配置详情

  • harness_arc_challenge_25

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|arc:challenge|25_2023-07-24T12:48:39.011618.parquet
  • harness_drop_3

    • 分片:2023_09_23T04_14_08.122970, latest
    • 路径:**/details_harness|drop|3_2023-09-23T04-14-08.122970.parquet
  • harness_gsm8k_5

    • 分片:2023_09_23T04_14_08.122970, latest
    • 路径:**/details_harness|gsm8k|5_2023-09-23T04-14-08.122970.parquet
  • harness_hellaswag_10

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hellaswag|10_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:多个路径,例如 **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_abstract_algebra_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_anatomy_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_astronomy_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_business_ethics_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_clinical_knowledge_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_college_biology_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_college_chemistry_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_college_computer_science_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_college_mathematics_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_college_medicine_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_college_physics_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_computer_security_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_conceptual_physics_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_econometrics_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_electrical_engineering_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-24T12:48:39.011618.parquet
  • harness_hendrycksTest_elementary_mathematics_5

    • 分片:2023_07_24T12_48_39.011618, latest
    • 路径:**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-24T12:48:39.011618.parquet

以上是数据集的详细概述,包括数据集名称、描述、组成、加载示例、最新结果以及各个配置的详细信息。

二维码
社区交流群
二维码
科研交流群
商业服务