open-llm-leaderboard-old/details_sonthenguyen__OpenHermes-2.5-Mistral-7B-mt-bench-DPO
收藏数据集概述
数据集摘要
该数据集是在模型 sonthenguyen/OpenHermes-2.5-Mistral-7B-mt-bench-DPO 在 Open LLM Leaderboard 上的评估运行期间自动创建的。数据集由 63 个配置组成,每个配置对应一个评估任务。
数据集结构
数据集包含 1 次运行,每个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。
额外配置
有一个额外的配置 "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。
加载数据
可以使用以下代码加载特定运行的详细信息: python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_sonthenguyen__OpenHermes-2.5-Mistral-7B-mt-bench-DPO", "harness_winogrande_5", split="train")
最新结果
以下是 最新结果 的摘要: python { "all": { "acc": 0.639538124544364, "acc_stderr": 0.032278452644941634, "acc_norm": 0.6420765688344944, "acc_norm_stderr": 0.0329201960227087, "mc1": 0.3598531211750306, "mc1_stderr": 0.016801860466677157, "mc2": 0.5290826380707278, "mc2_stderr": 0.015286081575472645 }, ... }
配置详情
-
config_name: harness_arc_challenge_25
- data_files:
- split: 2024_02_04T07_27_05.423264
- path: **/details_harness|arc:challenge|25_2024-02-04T07-27-05.423264.parquet
- split: latest
- path: **/details_harness|arc:challenge|25_2024-02-04T07-27-05.423264.parquet
- split: 2024_02_04T07_27_05.423264
- data_files:
-
config_name: harness_gsm8k_5
- data_files:
- split: 2024_02_04T07_27_05.423264
- path: **/details_harness|gsm8k|5_2024-02-04T07-27-05.423264.parquet
- split: latest
- path: **/details_harness|gsm8k|5_2024-02-04T07-27-05.423264.parquet
- split: 2024_02_04T07_27_05.423264
- data_files:
-
config_name: harness_hellaswag_10
- data_files:
- split: 2024_02_04T07_27_05.423264
- path: **/details_harness|hellaswag|10_2024-02-04T07-27-05.423264.parquet
- split: latest
- path: **/details_harness|hellaswag|10_2024-02-04T07-27-05.423264.parquet
- split: 2024_02_04T07_27_05.423264
- data_files:
-
config_name: harness_hendrycksTest_5
- data_files:
- split: 2024_02_04T07_27_05.423264
- path:
- **/details_harness|hendrycksTest-abstract_algebra|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-anatomy|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-astronomy|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-business_ethics|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-clinical_knowledge|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-college_biology|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-college_chemistry|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-college_computer_science|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-college_mathematics|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-college_medicine|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-college_physics|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-computer_security|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-conceptual_physics|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-econometrics|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-electrical_engineering|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-elementary_mathematics|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-formal_logic|5_2024-02-04T07-27-05.423264.parquet
- **/details_harness|hendrycksTest-global_facts|5_2024-02-04T07-27-05.423264.parquet }
- path:
- split: 2024_02_04T07_27_05.423264
- data_files:



