open-llm-leaderboard-old/details_Henk717__airochronos-33B
收藏Hugging Face2023-09-17 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_Henk717__airochronos-33B
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是在模型Henk717/airochronos-33B在Open LLM Leaderboard上评估运行期间自动创建的。它由64个配置组成,每个配置对应一个评估任务。数据集是从4次运行中创建的,每次运行作为每个配置中的一个特定分割,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还提供了如何使用Python代码加载运行细节的示例。
该数据集是在模型Henk717/airochronos-33B在Open LLM Leaderboard上评估运行期间自动创建的。它由64个配置组成,每个配置对应一个评估任务。数据集是从4次运行中创建的,每次运行作为每个配置中的一个特定分割,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还提供了如何使用Python代码加载运行细节的示例。
提供机构:
open-llm-leaderboard-old
原始信息汇总
数据集概述
数据集创建
- 数据集是在对模型 Henk717/airochronos-33B 进行评估运行期间自动创建的。
- 该数据集用于 Open LLM Leaderboard 上的评估。
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。
- 数据集由 4 次运行创建,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。
结果配置
- 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
- 加载特定运行的详细信息示例: python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_Henk717__airochronos-33B", "harness_winogrande_5", split="train")
最新结果
- 最新结果来自 2023-09-17T22:07:20.672645 运行,包含多个任务的结果: python { "all": { "em": 0.003145973154362416, "em_stderr": 0.0005734993648436351, "f1": 0.06925440436241624, "f1_stderr": 0.0014771385536763682, "acc": 0.46521874156655235, "acc_stderr": 0.010430187536918111 }, "harness|drop|3": { "em": 0.003145973154362416, "em_stderr": 0.0005734993648436351, "f1": 0.06925440436241624, "f1_stderr": 0.0014771385536763682 }, "harness|gsm8k|5": { "acc": 0.1372251705837756, "acc_stderr": 0.009477808244600422 }, "harness|winogrande|5": { "acc": 0.7932123125493291, "acc_stderr": 0.011382566829235798 } }
配置详情
- 数据集包含多个配置,每个配置对应不同的任务和运行时间戳。以下是部分配置示例:
harness_arc_challenge_25harness_drop_3harness_gsm8k_5harness_hellaswag_10harness_hendrycksTest_5harness_hendrycksTest_abstract_algebra_5harness_hendrycksTest_anatomy_5harness_hendrycksTest_astronomy_5harness_hendrycksTest_business_ethics_5harness_hendrycksTest_clinical_knowledge_5harness_hendrycksTest_college_biology_5harness_hendrycksTest_college_chemistry_5harness_hendrycksTest_college_computer_science_5harness_hendrycksTest_college_mathematics_5harness_hendrycksTest_college_medicine_5harness_hendrycksTest_college_physics_5harness_hendrycksTest_computer_security_5harness_hendrycksTest_conceptual_physics_5harness_hendrycksTest_econometrics_5harness_hendrycksTest_electrical_engineering_5
数据文件路径
- 每个配置包含多个数据文件,每个文件对应不同的分割和运行时间戳。例如:
harness_arc_challenge_25配置包含以下数据文件路径:**/details_harness|arc:challenge|25_2023-08-17T12:26:49.704789.parquet
harness_drop_3配置包含以下数据文件路径:**/details_harness|drop|3_2023-09-16T22-55-10.209177.parquet**/details_harness|drop|3_2023-09-17T00-16-43.512970.parquet**/details_harness|drop|3_2023-09-17T22-07-20.672645.parquet
harness_gsm8k_5配置包含以下数据文件路径:**/details_harness|gsm8k|5_2023-09-16T22-55-10.209177.parquet**/details_harness|gsm8k|5_2023-09-17T00-16-43.512970.parquet**/details_harness|gsm8k|5_2023-09-17T22-07-20.672645.parquet
harness_hellaswag_10配置包含以下数据文件路径:**/details_harness|hellaswag|10_2023-08-17T12:26:49.704789.parquet
harness_hendrycksTest_5配置包含以下数据文件路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-anatomy|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-astronomy|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-business_ethics|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-college_biology|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-college_chemistry|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-college_computer_science|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-college_mathematics|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-college_medicine|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-college_physics|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-computer_security|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-econometrics|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-formal_logic|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-global_facts|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-high_school_biology|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-high_school_geography|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-high_school_physics|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-high_school_psychology|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-high_school_statistics|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-high_school_us_history|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-high_school_world_history|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-human_aging|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-human_sexuality|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-international_law|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-jurisprudence|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-logical_fallacies|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-machine_learning|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-management|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-marketing|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-medical_genetics|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-miscellaneous|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-moral_disputes|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-moral_scenarios|5_2023-08-17T12:26:49.704789.parquet**/details_harness|hendrycksTest-nutrition|5_2023-08-17T12:26:49.704789.parquet- `**/details_harness|hendrycksTest-philosophy|5_2023-08-1



