five

open-llm-leaderboard/details_OpenAssistant__pythia-12b-sft-v8-2.5k-steps

收藏
Hugging Face2023-10-19 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_OpenAssistant__pythia-12b-sft-v8-2.5k-steps
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OpenAssistant/pythia-12b-sft-v8-2.5k-steps进行评估时自动生成的。数据集包含64个配置,每个配置对应一个被评估的任务。数据集由2次运行生成,每次运行在每个配置中表示为特定的分割,分割名称由运行的时间戳命名。train分割始终指向最新的结果。此外,一个名为results的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Python代码加载运行细节的示例,并包含了特定运行的最新结果。

This dataset was automatically generated when evaluating the model OpenAssistant/pythia-12b-sft-v8-2.5k-steps on the Open LLM Leaderboard. It consists of 64 configurations, with each configuration corresponding to one evaluated task. The dataset is generated from two independent runs, where each run is represented as a dedicated split under each configuration, and the split name is named after the timestamp of the corresponding run. The "train" split always points to the most recent evaluation results. Furthermore, a configuration named "results" stores the aggregated results across all runs, which are utilized to compute and display the aggregated metrics on the Open LLM Leaderboard. The accompanying README also provides examples of loading run details via Python code, and includes the latest results from specific runs.
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 OpenAssistant/pythia-12b-sft-v8-2.5k-steps

数据集来源

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找到一个特定的分割,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_OpenAssistant__pythia-12b-sft-v8-2.5k-steps", "harness_winogrande_5", split="train")

最新结果

  • 以下是 2023-10-19T06:05:53.274569 运行的最新结果: python { "all": { "em": 0.0008389261744966443, "em_stderr": 0.0002964962989801232, "f1": 0.04918729026845652, "f1_stderr": 0.0011958323498480873, "acc": 0.3760981059411563, "acc_stderr": 0.010720714478256874 }, "harness|drop|3": { "em": 0.0008389261744966443, "em_stderr": 0.0002964962989801232, "f1": 0.04918729026845652, "f1_stderr": 0.0011958323498480873 }, "harness|gsm8k|5": { "acc": 0.09552691432903715, "acc_stderr": 0.008096605771155738 }, "harness|winogrande|5": { "acc": 0.6566692975532754, "acc_stderr": 0.013344823185358009 } }

配置详情

  • harness_arc_challenge_25

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|arc:challenge|25_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|arc:challenge|25_2023-07-19T18:14:20.845496.parquet
  • harness_drop_3

    • 分割:2023_10_19T06_05_53.274569
    • 路径:**/details_harness|drop|3_2023-10-19T06-05-53.274569.parquet
    • 分割:latest
    • 路径:**/details_harness|drop|3_2023-10-19T06-05-53.274569.parquet
  • harness_gsm8k_5

    • 分割:2023_10_19T06_05_53.274569
    • 路径:**/details_harness|gsm8k|5_2023-10-19T06-05-53.274569.parquet
    • 分割:latest
    • 路径:**/details_harness|gsm8k|5_2023-10-19T06-05-53.274569.parquet
  • harness_hellaswag_10

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hellaswag|10_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|hellaswag|10_2023-07-19T18:14:20.845496.parquet
  • harness_hendrycksTest_5

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:14:20.845496.parquet 等 36 个文件
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:14:20.845496.parquet 等 36 个文件
  • harness_hendrycksTest_abstract_algebra_5

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:14:20.845496.parquet
  • harness_hendrycksTest_anatomy_5

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-07-19T18:14:20.845496.parquet
  • harness_hendrycksTest_astronomy_5

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-07-19T18:14:20.845496.parquet
  • harness_hendrycksTest_business_ethics_5

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T18:14:20.845496.parquet
  • harness_hendrycksTest_clinical_knowledge_5

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T18:14:20.845496.parquet
  • harness_hendrycksTest_college_biology_5

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-07-19T18:14:20.845496.parquet
  • harness_hendrycksTest_college_chemistry_5

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T18:14:20.845496.parquet
  • harness_hendrycksTest_college_computer_science_5

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T18:14:20.845496.parquet
  • harness_hendrycksTest_college_mathematics_5

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T18:14:20.845496.parquet
  • harness_hendrycksTest_college_medicine_5

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T18:14:20.845496.parquet
  • harness_hendrycksTest_college_physics_5

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-07-19T18:14:20.845496.parquet
  • harness_hendrycksTest_computer_security_5

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-07-19T18:14:20.845496.parquet
  • harness_hendrycksTest_conceptual_physics_5

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T18:14:20.845496.parquet
  • harness_hendrycksTest_econometrics_5

    • 分割:2023_07_19T18_14_20.845496
    • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-07-19T18:14:20.845496.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-07-19T18:14:20.845496.parquet
  • harness_hendrycksTest_electrical_engineering_5

    • 分割:2023_07_19T18_14_20.84
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作