open-llm-leaderboard-old/details_pmking27__PrathameshLLM-2B
收藏Hugging Face2024-04-07 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_pmking27__PrathameshLLM-2B
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pmking27/PrathameshLLM-2B进行评估时自动生成的。数据集包含63个配置,每个配置对应一个评估任务。数据集由一次运行的记录组成,每次运行在每个配置中作为一个特定的分割,分割名称使用运行的时间戳。train分割始终指向最新的结果。此外,还有一个results配置,存储了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。README还提供了一个Python代码片段来加载数据集,并列出了特定运行的最新结果。
该数据集是在Open LLM Leaderboard上对模型pmking27/PrathameshLLM-2B进行评估时自动生成的。数据集包含63个配置,每个配置对应一个评估任务。数据集由一次运行的记录组成,每次运行在每个配置中作为一个特定的分割,分割名称使用运行的时间戳。train分割始终指向最新的结果。此外,还有一个results配置,存储了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。README还提供了一个Python代码片段来加载数据集,并列出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
数据集创建
- 创建背景:该数据集是在对模型 pmking27/PrathameshLLM-2B 进行评估运行时自动创建的。
- 评估平台:数据集的创建与 Open LLM Leaderboard 相关联。
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。每个运行结果作为一个特定的分割(split)存储在每个配置中,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。
- 汇总结果:一个额外的配置 "results" 存储所有运行的汇总结果,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_pmking27__PrathameshLLM-2B", "harness_winogrande_5", split="train")
最新结果
- 最新结果时间戳:2024-04-07T15:12:15.779927
- 结果示例: python { "all": { "acc": 0.38448642166678587, "acc_stderr": 0.03412711045658374, "acc_norm": 0.3882749361001093, "acc_norm_stderr": 0.03491709236659397, "mc1": 0.2913096695226438, "mc1_stderr": 0.01590598704818483, "mc2": 0.44689893094855493, "mc2_stderr": 0.014502506509597363 }, "harness|arc:challenge|25": { "acc": 0.4257679180887372, "acc_stderr": 0.014449464278868809, "acc_norm": 0.447098976109215, "acc_norm_stderr": 0.014529380160526845 }, "harness|hellaswag|10": { "acc": 0.5071698864767975, "acc_stderr": 0.00498926836296872, "acc_norm": 0.6840270862378013, "acc_norm_stderr": 0.004639520453444027 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.26, "acc_stderr": 0.04408440022768081, "acc_norm": 0.26, "acc_norm_stderr": 0.04408440022768081 }, # 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_07T15_12_15.779927
- 路径:
**/details_harness|arc:challenge|25_2024-04-07T15-12-15.779927.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|arc:challenge|25_2024-04-07T15-12-15.779927.parquet
- 路径:
- 分割:2024_04_07T15_12_15.779927
- 数据文件:
-
配置名称:harness_gsm8k_5
- 数据文件:
- 分割:2024_04_07T15_12_15.779927
- 路径:
**/details_harness|gsm8k|5_2024-04-07T15-12-15.779927.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|gsm8k|5_2024-04-07T15-12-15.779927.parquet
- 路径:
- 分割:2024_04_07T15_12_15.779927
- 数据文件:
-
配置名称:harness_hellaswag_10
- 数据文件:
- 分割:2024_04_07T15_12_15.779927
- 路径:
**/details_harness|hellaswag|10_2024-04-07T15-12-15.779927.parquet
- 路径:
- 分割:latest
- 路径:
**/details_harness|hellaswag|10_2024-04-07T15-12-15.779927.parquet
- 路径:
- 分割:2024_04_07T15_12_15.779927
- 数据文件:
-
配置名称:harness_hendrycksTest_5
- 数据文件:
- 分割:2024_04_07T15_12_15.779927
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-anatomy|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-astronomy|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-business_ethics|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-college_biology|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-college_chemistry|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-college_computer_science|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-college_mathematics|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-college_medicine|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-college_physics|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-computer_security|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-econometrics|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-formal_logic|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-global_facts|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-high_school_biology|5_2024-04-07T15-12-15.779927.parquet
其他任务的路径...
- 路径:
- 分割:latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-anatomy|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-astronomy|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-business_ethics|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-college_biology|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-college_chemistry|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-college_computer_science|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-college_mathematics|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-college_medicine|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-college_physics|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-computer_security|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-econometrics|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-formal_logic|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-global_facts|5_2024-04-07T15-12-15.779927.parquet**/details_harness|hendrycksTest-high_school_biology|5_2024-04-07T15-12-15.779927.parquet
其他任务的路径...
- 路径:
- 分割:2024_04_07T15_12_15.779927
- 数据文件:



