open-llm-leaderboard-old/details_KnutJaegersberg__Deita-32b
收藏Hugging Face2024-04-08 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_KnutJaegersberg__Deita-32b
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KnutJaegersberg/Deita-32b的评估运行期间自动创建的,用于Open LLM Leaderboard的评估。数据集由63个配置组成,每个配置对应一个评估任务。数据集从1次运行中创建,每次运行都可以在特定配置中找到,分割以运行的时间戳命名。train分割始终指向最新结果。此外,results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。
该数据集是在模型KnutJaegersberg/Deita-32b的评估运行期间自动创建的,用于Open LLM Leaderboard的评估。数据集由63个配置组成,每个配置对应一个评估任务。数据集从1次运行中创建,每次运行都可以在特定配置中找到,分割以运行的时间戳命名。train分割始终指向最新结果。此外,results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。
提供机构:
open-llm-leaderboard-old
原始信息汇总
数据集概述
数据集创建
- 创建背景:该数据集是在对模型 KnutJaegersberg/Deita-32b 进行评估运行期间自动创建的。
- 评估平台:数据集用于 Open LLM Leaderboard。
数据集结构
- 配置数量:数据集包含 63 个配置,每个配置对应一个评估任务。
- 运行次数:数据集从 1 次运行中创建。每个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。
- 最新结果:"train" 分割始终指向最新的结果。
- 聚合结果:一个额外的配置 "results" 存储所有运行的聚合结果,用于计算和显示在 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_KnutJaegersberg__Deita-32b", "harness_winogrande_5", split="train")
最新结果
- 最新结果时间戳:2024-04-08T14:30:53.438212
- 结果示例: python { "all": { "acc": 0.736367854664048, "acc_stderr": 0.029323294534426905, "acc_norm": 0.7394575954204685, "acc_norm_stderr": 0.029896267973729986, "mc1": 0.408812729498164, "mc1_stderr": 0.01720995215164173, "mc2": 0.5811395725483436, "mc2_stderr": 0.014925357411175256 }, "harness|arc:challenge|25": { "acc": 0.5861774744027304, "acc_stderr": 0.01439273000922101, "acc_norm": 0.6305460750853242, "acc_norm_stderr": 0.01410457836649189 }, "harness|hellaswag|10": { "acc": 0.6500697072296355, "acc_stderr": 0.004759729267943188, "acc_norm": 0.8467436765584545, "acc_norm_stderr": 0.003594981823319894 }, # 其他任务的结果... }
配置详情
-
配置名称:harness_arc_challenge_25
- 数据文件:
- 分割:2024_04_08T14_30_53.438212
- 路径:
**/details_harness|arc:challenge|25_2024-04-08T14-30-53.438212.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|arc:challenge|25_2024-04-08T14-30-53.438212.parquet
- 路径:
- 分割:2024_04_08T14_30_53.438212
- 数据文件:
-
配置名称:harness_gsm8k_5
- 数据文件:
- 分割:2024_04_08T14_30_53.438212
- 路径:
**/details_harness|gsm8k|5_2024-04-08T14-30-53.438212.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|gsm8k|5_2024-04-08T14-30-53.438212.parquet
- 路径:
- 分割:2024_04_08T14_30_53.438212
- 数据文件:
-
配置名称:harness_hellaswag_10
- 数据文件:
- 分割:2024_04_08T14_30_53.438212
- 路径:
**/details_harness|hellaswag|10_2024-04-08T14-30-53.438212.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|hellaswag|10_2024-04-08T14-30-53.438212.parquet
- 路径:
- 分割:2024_04_08T14_30_53.438212
- 数据文件:
-
配置名称:harness_hendrycksTest_5
- 数据文件:
- 分割:2024_04_08T14_30_53.438212
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-anatomy|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-astronomy|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-business_ethics|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-college_biology|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-college_chemistry|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-college_computer_science|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-college_mathematics|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-college_medicine|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-college_physics|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-computer_security|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-econometrics|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-formal_logic|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-global_facts|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-high_school_biology|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-high_school_chemistry|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-high_school_computer_science|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-high_school_european_history|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-high_school_geography|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-high_school_government_and_politics|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-high_school_macroeconomics|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-high_school_mathematics|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-high_school_microeconomics|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-high_school_physics|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-high_school_psychology|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-high_school_statistics|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-high_school_us_history|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-high_school_world_history|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-human_aging|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-human_sexuality|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-international_law|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-jurisprudence|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-logical_fallacies|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-machine_learning|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-management|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-marketing|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-medical_genetics|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-miscellaneous|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-moral_disputes|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-moral_scenarios|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-nutrition|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-philosophy|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-prehistory|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-professional_accounting|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-professional_law|5_2024-04-08T14-30-53.438212.parquet**/details_harness|hendrycksTest-professional_medicine|5_2024-04-08T14-30-53.438212.parquet- `**/details_harness|hendrycksTest-professional_psychology|5_2024-04-08T14-30-5
- 路径:
- 分割:2024_04_08T14_30_53.438212
- 数据文件:



