five

open-llm-leaderboard/details_KoboldAI__OPT-6.7B-Nerybus-Mix

收藏
Hugging Face2023-10-15 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_KoboldAI__OPT-6.7B-Nerybus-Mix
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KoboldAI/OPT-6.7B-Nerybus-Mix进行评估运行时自动创建的。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次运行创建,每次运行在每个配置中表示为特定的分割。train分割始终指向最新的结果。一个名为results的额外配置存储了所有运行的聚合结果,这些结果用于计算和显示Open LLM Leaderboard上的聚合指标。README还提供了如何使用`datasets`库中的`load_dataset`函数加载运行中的详细信息的示例。

This dataset was automatically created during the evaluation run of the model KoboldAI/OPT-6.7B-Nerybus-Mix on the Open LLM Leaderboard. The dataset consists of 64 configurations, each corresponding to one evaluation task. The dataset is constructed from two evaluation runs, where each run is represented as a specific split under each configuration. The `train` split always points to the most recent results. An additional configuration named `results` stores the aggregated results across all runs, which are used to calculate and display the aggregate metrics on the Open LLM Leaderboard. The README also provides examples of how to use the `load_dataset` function from the `datasets` library to load detailed information about the runs.
提供机构:
open-llm-leaderboard
原始信息汇总

数据集概述

数据集简介

该数据集是在对模型 KoboldAI/OPT-6.7B-Nerybus-Mix 进行评估运行期间自动创建的,用于 Open LLM Leaderboard

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。
  • 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_KoboldAI__OPT-6.7B-Nerybus-Mix", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-15T02:15:38.623637 运行的最新结果: python { "all": { "em": 0.001153523489932886, "em_stderr": 0.0003476179896857095, "f1": 0.04849203020134232, "f1_stderr": 0.0011914526339428257, "acc": 0.3293629684060238, "acc_stderr": 0.007892170416107272 }, "harness|drop|3": { "em": 0.001153523489932886, "em_stderr": 0.0003476179896857095, "f1": 0.04849203020134232, "f1_stderr": 0.0011914526339428257 }, "harness|gsm8k|5": { "acc": 0.0075815011372251705, "acc_stderr": 0.0023892815120772222 }, "harness|winogrande|5": { "acc": 0.6511444356748224, "acc_stderr": 0.013395059320137324 } }

配置详情

以下是部分配置及其数据文件路径:

  • config_name: harness_arc_challenge_25

    • split: 2023_07_19T17_22_17.446563
      • path: **/details_harness|arc:challenge|25_2023-07-19T17:22:17.446563.parquet
    • split: latest
      • path: **/details_harness|arc:challenge|25_2023-07-19T17:22:17.446563.parquet
  • config_name: harness_drop_3

    • split: 2023_10_15T02_15_38.623637
      • path: **/details_harness|drop|3_2023-10-15T02-15-38.623637.parquet
    • split: latest
      • path: **/details_harness|drop|3_2023-10-15T02-15-38.623637.parquet
  • config_name: harness_gsm8k_5

    • split: 2023_10_15T02_15_38.623637
      • path: **/details_harness|gsm8k|5_2023-10-15T02-15-38.623637.parquet
    • split: latest
      • path: **/details_harness|gsm8k|5_2023-10-15T02-15-38.623637.parquet
  • config_name: harness_hellaswag_10

    • split: 2023_07_19T17_22_17.446563
      • path: **/details_harness|hellaswag|10_2023-07-19T17:22:17.446563.parquet
    • split: latest
      • path: **/details_harness|hellaswag|10_2023-07-19T17:22:17.446563.parquet
  • config_name: harness_hendrycksTest_5

    • split: 2023_07_19T17_22_17.446563
      • path:
        • **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-high_school_physics|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-high_school_psychology|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-high_school_statistics|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-high_school_us_history|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-high_school_world_history|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-human_aging|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-human_sexuality|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-international_law|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-jurisprudence|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-logical_fallacies|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-machine_learning|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-management|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-marketing|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-medical_genetics|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-miscellaneous|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-moral_disputes|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-moral_scenarios|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-nutrition|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-philosophy|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-prehistory|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-professional_accounting|5_2023-07-19T17:22:17.446563.parquet
        • **/details_harness|hendrycksTest-professional_law|5_
二维码
社区交流群
二维码
科研交流群
商业服务