open-llm-leaderboard-old/details_KoboldAI__OPT-6.7B-Erebus
收藏Hugging Face2023-10-22 更新2024-06-22 收录
下载链接：
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_KoboldAI__OPT-6.7B-Erebus
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KoboldAI/OPT-6.7B-Erebus模型进行评估时自动创建的。数据集由64个配置组成，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。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，用于在Open LLM Leaderboard上计算和显示聚合指标。文件还提供了如何使用Hugging Face datasets库加载数据集的示例。
该数据集是在Open LLM Leaderboard上对KoboldAI/OPT-6.7B-Erebus模型进行评估时自动创建的。数据集由64个配置组成，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。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，用于在Open LLM Leaderboard上计算和显示聚合指标。文件还提供了如何使用Hugging Face datasets库加载数据集的示例。
提供机构：
open-llm-leaderboard-old
原始信息汇总
数据集概述
数据集简介
该数据集是在对模型 KoboldAI/OPT-6.7B-Erebus 进行评估运行期间自动创建的，用于 Open LLM Leaderboard。
数据集结构
- 配置数量：64个配置，每个配置对应一个评估任务。
- 数据来源：数据集由2次运行结果组成，每次运行的结果可以在每个配置中找到，并以运行的时间戳命名。
- 最新结果："train" 分割总是指向最新的结果。
- 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_KoboldAI__OPT-6.7B-Erebus", "harness_winogrande_5", split="train")
最新结果
以下是 2023-10-22T01:42:09.506783 运行的最新结果： python { "all": { "em": 0.001153523489932886, "em_stderr": 0.0003476179896857095, "f1": 0.04858431208053701, "f1_stderr": 0.0011921353872508366, "acc": 0.33522023153796765, "acc_stderr": 0.00806896604240255 }, "harness|drop|3": { "em": 0.001153523489932886, "em_stderr": 0.0003476179896857095, "f1": 0.04858431208053701, "f1_stderr": 0.0011921353872508366 }, "harness|gsm8k|5": { "acc": 0.01061410159211524, "acc_stderr": 0.0028227133223877043 }, "harness|winogrande|5": { "acc": 0.65982636148382, "acc_stderr": 0.013315218762417397 } }
配置详情
以下是部分配置及其数据文件路径：
-
harness_arc_challenge_25
- 分割：2023_07_19T17_20_54.049241
- 路径：
**/details_harness|arc:challenge|25_2023-07-19T17:20:54.049241.parquet
- 路径：
- 分割：latest
- 路径：
**/details_harness|arc:challenge|25_2023-07-19T17:20:54.049241.parquet
- 路径：
- 分割：2023_07_19T17_20_54.049241
-
harness_drop_3
- 分割：2023_10_22T01_42_09.506783
- 路径：
**/details_harness|drop|3_2023-10-22T01-42-09.506783.parquet
- 路径：
- 分割：latest
- 路径：
**/details_harness|drop|3_2023-10-22T01-42-09.506783.parquet
- 路径：
- 分割：2023_10_22T01_42_09.506783
-
harness_gsm8k_5
- 分割：2023_10_22T01_42_09.506783
- 路径：
**/details_harness|gsm8k|5_2023-10-22T01-42-09.506783.parquet
- 路径：
- 分割：latest
- 路径：
**/details_harness|gsm8k|5_2023-10-22T01-42-09.506783.parquet
- 路径：
- 分割：2023_10_22T01_42_09.506783
-
harness_hellaswag_10
- 分割：2023_07_19T17_20_54.049241
- 路径：
**/details_harness|hellaswag|10_2023-07-19T17:20:54.049241.parquet
- 路径：
- 分割：latest
- 路径：
**/details_harness|hellaswag|10_2023-07-19T17:20:54.049241.parquet
- 路径：
- 分割：2023_07_19T17_20_54.049241
-
harness_hendrycksTest_5
- 分割：2023_07_19T17_20_54.049241
- 路径：
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-anatomy|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-astronomy|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-college_biology|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-college_physics|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-computer_security|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-econometrics|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-formal_logic|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-global_facts|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-high_school_physics|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-high_school_psychology|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-high_school_statistics|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-high_school_us_history|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-high_school_world_history|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-human_aging|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-human_sexuality|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-international_law|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-jurisprudence|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-logical_fallacies|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-machine_learning|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-management|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-marketing|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-medical_genetics|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-miscellaneous|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-moral_disputes|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-moral_scenarios|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-nutrition|5_2023-07-19T17:20:54.049241.parquet - 路径：
**/details_harness|hendrycksTest-philosophy|5_2023-07-19T17:20:54.049241.parquet - 路径：`**/details_har
- 路径：
- 分割：2023_07_19T17_20_54.049241
