five

open-llm-leaderboard/details_upstage__llama-30b-instruct-2048

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_upstage__llama-30b-instruct-2048
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upstage/llama-30b-instruct-2048进行评估时自动生成的。数据集包含64个配置,每个配置对应一个评估任务。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,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。README还提供了如何使用Hugging Face datasets库加载运行细节的示例。

This dataset was automatically generated during the evaluation of the model upstage/llama-30b-instruct-2048 on the Open LLM Leaderboard. It consists of 64 configurations, each corresponding to one evaluation task. The dataset is generated via two runs. Under each configuration, each run corresponds to a specific split, with the split’s name derived from the run’s timestamp. The train split always points to the most recent evaluation results. Additionally, 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 how to load run details using the Hugging Face Datasets library.
提供机构:
open-llm-leaderboard
原始信息汇总

数据集概述

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。

额外配置

  • 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_upstage__llama-30b-instruct-2048", "harness_winogrande_5", split="train")

最新结果

  • 最新结果来自 2023-10-19T00:52:48.467311 运行,包含以下指标: python { "all": { "em": 0.35539010067114096, "em_stderr": 0.004901633098104223, "f1": 0.44463611577181344, "f1_stderr": 0.004655171488287754, "acc": 0.48687269361101454, "acc_stderr": 0.010937111570073342 }, "harness|drop|3": { "em": 0.35539010067114096, "em_stderr": 0.004901633098104223, "f1": 0.44463611577181344, "f1_stderr": 0.004655171488287754 }, "harness|gsm8k|5": { "acc": 0.17816527672479152, "acc_stderr": 0.01054013252754947 }, "harness|winogrande|5": { "acc": 0.7955801104972375, "acc_stderr": 0.011334090612597212 } }

配置详情

  • harness_arc_challenge_25

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|arc:challenge|25_2023-07-19T12:29:43.161348.parquet
  • harness_drop_3

    • 分割:2023_10_19T00_52_48.467311, latest
    • 路径:**/details_harness|drop|3_2023-10-19T00-52-48.467311.parquet
  • harness_gsm8k_5

    • 分割:2023_10_19T00_52_48.467311, latest
    • 路径:**/details_harness|gsm8k|5_2023-10-19T00-52-48.467311.parquet
  • harness_hellaswag_10

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hellaswag|10_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:多个路径,例如 **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_abstract_algebra_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_anatomy_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_astronomy_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_business_ethics_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_clinical_knowledge_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_college_biology_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_college_chemistry_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_college_computer_science_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_college_mathematics_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_college_medicine_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_college_physics_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_computer_security_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_conceptual_physics_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_econometrics_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_electrical_engineering_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T12:29:43.161348.parquet
  • harness_hendrycksTest_elementary_mathematics_5

    • 分割:2023_07_19T12_29_43.161348, latest
    • 路径:**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T12:29:43.161348.parquet
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