five

open-llm-leaderboard-old/details_bigcode__starcoder

收藏
Hugging Face2024-02-14 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_bigcode__starcoder
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bigcode/starcoder进行评估时自动创建的。数据集由121个配置组成,每个配置对应一个评估任务。数据集由4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。此外,还有一个名为results的配置,存储了所有运行的聚合结果,用于计算和显示Open LLM Leaderboard上的聚合指标。README文件还提供了如何使用Python代码加载运行中的详细信息的示例,并提供了特定运行的最新结果。

该数据集是在Open LLM Leaderboard上对模型bigcode/starcoder进行评估时自动创建的。数据集由121个配置组成,每个配置对应一个评估任务。数据集由4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。此外,还有一个名为results的配置,存储了所有运行的聚合结果,用于计算和显示Open LLM Leaderboard上的聚合指标。README文件还提供了如何使用Python代码加载运行中的详细信息的示例,并提供了特定运行的最新结果。
提供机构:
open-llm-leaderboard-old
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 bigcode/starcoder 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 121 个配置,每个配置对应一个评估任务。
  • 数据集从 4 次运行中创建,每次运行在每个配置中作为一个特定的分割存在,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_bigcode__starcoder", "harness_winogrande_5", split="train")

最新结果

以下是 2024-02-14T22:50:56.838467 运行的最新结果

python { "all": { "acc": 0.2969189890806991, "acc_stderr": 0.03236365511067932, "acc_norm": 0.2979650690177265, "acc_norm_stderr": 0.033097159757475146, "mc1": 0.25091799265605874, "mc1_stderr": 0.015176985027707689, "mc2": 0.4130412207453783, "mc2_stderr": 0.014976467041499917 }, "harness|arc:challenge|25": { "acc": 0.28071672354948807, "acc_stderr": 0.013131238126975574, "acc_norm": 0.302901023890785, "acc_norm_stderr": 0.013428241573185349 }, "harness|hellaswag|10": { "acc": 0.37860983867755427, "acc_stderr": 0.004840493603166207, "acc_norm": 0.4787890858394742, "acc_norm_stderr": 0.004985289555586536 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.23, "acc_stderr": 0.042295258468165044, "acc_norm": 0.23, "acc_norm_stderr": 0.042295258468165044 }, # 其他任务的结果... }

配置详情

  • harness_arc_challenge_25

    • 分割:2023_08_28T09_53_59.312863, 2024_02_14T22_50_56.838467, latest
    • 路径:**/details_harness|arc:challenge|25_2023-08-28T09:53:59.312863.parquet, **/details_harness|arc:challenge|25_2024-02-14T22-50-56.838467.parquet
  • harness_gsm8k_5

    • 分割:2024_02_14T22_50_56.838467, latest
    • 路径:**/details_harness|gsm8k|5_2024-02-14T22-50-56.838467.parquet
  • harness_hellaswag_10

    • 分割:2023_08_28T09_53_59.312863, 2024_02_14T22_50_56.838467, latest
    • 路径:**/details_harness|hellaswag|10_2023-08-28T09:53:59.312863.parquet, **/details_harness|hellaswag|10_2024-02-14T22-50-56.838467.parquet
  • harness_hendrycksTest_5

    • 分割:2023_08_28T09_53_59.312863
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-anatomy|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-astronomy|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-business_ethics|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-college_biology|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-college_chemistry|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-college_computer_science|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-college_mathematics|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-college_medicine|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-college_physics|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-computer_security|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-conceptual_physics|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-econometrics|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-electrical_engineering|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-formal_logic|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-global_facts|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-high_school_biology|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-high_school_european_history|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-high_school_geography|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-high_school_physics|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-high_school_psychology|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-high_school_statistics|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-high_school_us_history|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-high_school_world_history|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-human_aging|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-human_sexuality|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-international_law|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-jurisprudence|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-logical_fallacies|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-machine_learning|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-management|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-marketing|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-medical_genetics|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-miscellaneous|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-moral_disputes|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-moral_scenarios|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-nutrition|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-philosophy|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-prehistory|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-professional_accounting|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-professional_law|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-professional_medicine|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-professional_psychology|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-public_relations|5_2023-08-28T09:53:59.312863.parquet, **/details_harness|hendrycksTest-security_studies|5_2023-08-28T09:53:59.312863.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