open-llm-leaderboard/details_lu-vae__llama2-13b-sharegpt4-test
收藏资源简介:
该数据集是在模型lu-vae/llama2-13b-sharegpt4-test的评估运行期间自动创建的,用于Open LLM Leaderboard。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,每次运行的结果作为特定配置中的一个分割,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,还有一个名为results的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。
This dataset was automatically generated during the evaluation run of the model lu-vae/llama2-13b-sharegpt4-test, and is designed for the Open LLM Leaderboard. It consists of 64 configurations, each corresponding to one evaluation task. The dataset is created from two runs, where the results of each run act as a dataset split under a specific configuration, with the split name being the timestamp of the corresponding run. The 'train' split always points to the most recent results. Additionally, there is a configuration named 'results' that stores the aggregated results from all runs, which is used to calculate and display aggregate metrics on the Open LLM Leaderboard.
数据集概述
数据集名为 "Evaluation run of lu-vae/llama2-13b-sharegpt4-test",是在模型 lu-vae/llama2-13b-sharegpt4-test 在 Open LLM Leaderboard 上的评估运行期间自动创建的。
数据集组成
- 数据集包含 64 个配置,每个配置对应一个评估任务。
- 数据集从 2 次运行中创建,每个运行可以在每个配置中作为一个特定的分片找到,分片名称使用运行的时间戳。
- "train" 分片总是指向最新的结果。
- 额外的配置 "results" 存储所有运行结果的聚合数据,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_lu-vae__llama2-13b-sharegpt4-test", "harness_winogrande_5", split="train")
最新结果
以下是 2023-10-25T11:44:44.255542 运行 的最新结果:
python { "all": { "em": 0.002202181208053691, "em_stderr": 0.00048005108166192744, "f1": 0.06610843120805385, "f1_stderr": 0.0014639889801036593, "acc": 0.44600618846762125, "acc_stderr": 0.01064352054021588 }, "harness|drop|3": { "em": 0.002202181208053691, "em_stderr": 0.00048005108166192744, "f1": 0.06610843120805385, "f1_stderr": 0.0014639889801036593 }, "harness|gsm8k|5": { "acc": 0.13115996967399546, "acc_stderr": 0.00929849923558786 }, "harness|winogrande|5": { "acc": 0.760852407261247, "acc_stderr": 0.011988541844843902 } }
配置详情
-
harness_arc_challenge_25
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|arc:challenge|25_2023-09-12T16-41-26.998548.parquet
-
harness_drop_3
- 分片:2023_10_25T11_44_44.255542, latest
- 路径:
**/details_harness|drop|3_2023-10-25T11-44-44.255542.parquet
-
harness_gsm8k_5
- 分片:2023_10_25T11_44_44.255542, latest
- 路径:
**/details_harness|gsm8k|5_2023-10-25T11-44-44.255542.parquet
-
harness_hellaswag_10
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hellaswag|10_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:多个路径,如
**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-12T16-41-26.998548.parquet等
-
harness_hendrycksTest_abstract_algebra_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_anatomy_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_astronomy_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_business_ethics_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_clinical_knowledge_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_college_biology_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_college_chemistry_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_college_computer_science_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_college_mathematics_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_college_medicine_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_college_physics_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_computer_security_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_conceptual_physics_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_econometrics_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_electrical_engineering_5
- 分片:2023_09_12T16_41_26.998548, latest
- 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-09-12T16-41-26.998548.parquet
-
harness_hendrycksTest_elementary_mathematics_5
- 分片:2023_09_12T16_41_26.998548
- 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-12T16-41-26.998548.parquet
以上是数据集的详细概述,包括数据集的组成、加载示例、最新结果以及各个配置的详细信息。



