open-llm-leaderboard-old/details_Swisslex__Mixtral-8x7b-DPO-v0.2
收藏数据集概述
数据集简介
该数据集是在模型Swisslex/Mixtral-8x7b-DPO-v0.2的评估运行期间自动创建的,用于Open LLM Leaderboard。
数据集组成
- 数据集包含63个配置,每个配置对应一个评估任务。
- 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。
- "train"分割总是指向最新的结果。
- 额外的"results"配置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。
数据加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_Swisslex__Mixtral-8x7b-DPO-v0.2", "harness_winogrande_5", split="train")
最新结果
以下是最新结果的摘要: python { "all": { "acc": 0.7090305152663554, "acc_stderr": 0.030408745551927793, "acc_norm": 0.7130050743913563, "acc_norm_stderr": 0.03099893935186279, "mc1": 0.4394124847001224, "mc1_stderr": 0.017374520482513707, "mc2": 0.5869091785507645, "mc2_stderr": 0.01561392560307738 }, "harness|arc:challenge|25": { "acc": 0.6715017064846417, "acc_stderr": 0.013724978465537293, "acc_norm": 0.7039249146757679, "acc_norm_stderr": 0.01334091608524626 }, "harness|hellaswag|10": { "acc": 0.6922923720374428, "acc_stderr": 0.004606015773125625, "acc_norm": 0.8773152758414658, "acc_norm_stderr": 0.0032740447231806155 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.36, "acc_stderr": 0.04824181513244218, "acc_norm": 0.36, "acc_norm_stderr": 0.04824181513244218 }, # 其他任务的结果... }
配置详情
-
harness_arc_challenge_25
- 分割:
2024_01_16T18_18_43.502951,latest - 路径:
**/details_harness|arc:challenge|25_2024-01-16T18-18-43.502951.parquet
- 分割:
-
harness_gsm8k_5
- 分割:
2024_01_16T18_18_43.502951,latest - 路径:
**/details_harness|gsm8k|5_2024-01-16T18-18-43.502951.parquet
- 分割:
-
harness_hellaswag_10
- 分割:
2024_01_16T18_18_43.502951,latest - 路径:
**/details_harness|hellaswag|10_2024-01-16T18-18-43.502951.parquet
- 分割:
-
harness_hendrycksTest_5
- 分割:
2024_01_16T18_18_43.502951,latest - 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-anatomy|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-astronomy|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-business_ethics|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-college_biology|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-college_chemistry|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-college_computer_science|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-college_mathematics|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-college_medicine|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-college_physics|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-computer_security|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-econometrics|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-formal_logic|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-global_facts|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-high_school_biology|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-high_school_chemistry|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-high_school_computer_science|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-high_school_european_history|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-high_school_geography|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-high_school_government_and_politics|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-high_school_macroeconomics|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-high_school_mathematics|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-high_school_microeconomics|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-high_school_physics|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-high_school_psychology|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-high_school_statistics|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-high_school_us_history|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-high_school_world_history|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-human_aging|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-human_sexuality|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-international_law|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-jurisprudence|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-logical_fallacies|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-machine_learning|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-management|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-marketing|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-medical_genetics|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-miscellaneous|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-moral_disputes|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-moral_scenarios|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-nutrition|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-philosophy|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-prehistory|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-professional_accounting|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-professional_law|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-professional_medicine|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-professional_psychology|5_2024-01-16T18-18-43.502951.parquet**/details_harness|hendrycksTest-public_relations|5_2024-01-16T18-18-43.502951.parquet- `**/details_harness|
- 分割:



