open-llm-leaderboard-old/details_meta-llama__Meta-Llama-3-70B
收藏数据集概述
数据集摘要
该数据集是在对模型 meta-llama/Meta-Llama-3-70B 进行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" 分片始终指向最新的结果。
- 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。
数据加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_meta-llama__Meta-Llama-3-70B_private", "harness_winogrande_5", split="train")
最新结果
这些是最新结果的示例,来自 2024-04-21T13:09:06.084236 的运行: python { "all": { "acc": 0.7890317570087406, "acc_stderr": 0.026707462510455523, "acc_norm": 0.7920388208739257, "acc_norm_stderr": 0.027217966289275614, "mc1": 0.29865361077111385, "mc1_stderr": 0.01602157061376854, "mc2": 0.45562368201500897, "mc2_stderr": 0.013963421323817822 }, "harness|arc:challenge|25": { "acc": 0.6510238907849829, "acc_stderr": 0.0139289334613825, "acc_norm": 0.6877133105802048, "acc_norm_stderr": 0.013542598541688067 }, "harness|hellaswag|10": { "acc": 0.6951802429794861, "acc_stderr": 0.004593902601979336, "acc_norm": 0.8798048197570205, "acc_norm_stderr": 0.003245250394565294 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.48, "acc_stderr": 0.05021167315686779, "acc_norm": 0.48, "acc_norm_stderr": 0.05021167315686779 }, # 其他任务的结果... }
配置详情
-
config_name: harness_arc_challenge_25
- split: 2024_04_21T13_09_06.084236
- path: **/details_harness|arc:challenge|25_2024-04-21T13-09-06.084236.parquet
- split: latest
- path: **/details_harness|arc:challenge|25_2024-04-21T13-09-06.084236.parquet
- split: 2024_04_21T13_09_06.084236
-
config_name: harness_gsm8k_5
- split: 2024_04_21T13_09_06.084236
- path: **/details_harness|gsm8k|5_2024-04-21T13-09-06.084236.parquet
- split: latest
- path: **/details_harness|gsm8k|5_2024-04-21T13-09-06.084236.parquet
- split: 2024_04_21T13_09_06.084236
-
config_name: harness_hellaswag_10
- split: 2024_04_21T13_09_06.084236
- path: **/details_harness|hellaswag|10_2024-04-21T13-09-06.084236.parquet
- split: latest
- path: **/details_harness|hellaswag|10_2024-04-21T13-09-06.084236.parquet
- split: 2024_04_21T13_09_06.084236
-
config_name: harness_hendrycksTest_5
- split: 2024_04_21T13_09_06.084236
- path:
- **/details_harness|hendrycksTest-abstract_algebra|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-anatomy|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-astronomy|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-business_ethics|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-clinical_knowledge|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-college_biology|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-college_chemistry|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-college_computer_science|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-college_mathematics|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-college_medicine|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-college_physics|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-computer_security|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-conceptual_physics|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-econometrics|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-electrical_engineering|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-elementary_mathematics|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-formal_logic|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-global_facts|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-high_school_biology|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-high_school_chemistry|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-high_school_computer_science|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-high_school_european_history|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-high_school_geography|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-high_school_government_and_politics|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-high_school_macroeconomics|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-high_school_mathematics|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-high_school_microeconomics|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-high_school_physics|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-high_school_psychology|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-high_school_statistics|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-high_school_us_history|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-high_school_world_history|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-human_aging|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-human_sexuality|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-international_law|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-jurisprudence|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-logical_fallacies|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-machine_learning|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-management|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-marketing|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-medical_genetics|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-miscellaneous|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-moral_disputes|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-moral_scenarios|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-nutrition|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-philosophy|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-prehistory|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-professional_accounting|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-professional_law|5_2024-04-21T13-09-06.084236.parquet
- **/details_harness|hendrycksTest-professional_medicine|5_2024-04-21T13-09-06.084236.parquet
- **/
- path:
- split: 2024_04_21T13_09_06.084236



