open-llm-leaderboard/details_OpenBuddy__openbuddy-llama2-13b-v11.1-bf16
收藏数据集概述
数据集简介
该数据集是在评估模型OpenBuddy/openbuddy-llama2-13b-v11.1-bf16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3次运行中创建,每次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_OpenBuddy__openbuddy-llama2-13b-v11.1-bf16", "harness_winogrande_5", split="train")
最新结果
以下是最新结果(来自2023-10-15T18:23:36.599949的运行): python { "all": { "em": 0.3735318791946309, "em_stderr": 0.004953965475225372, "f1": 0.4288569630872499, "f1_stderr": 0.004814754523733826, "acc": 0.48908640816959104, "acc_stderr": 0.012113244925946991 }, "harness|drop|3": { "em": 0.3735318791946309, "em_stderr": 0.004953965475225372, "f1": 0.4288569630872499, "f1_stderr": 0.004814754523733826 }, "harness|gsm8k|5": { "acc": 0.24336618650492797, "acc_stderr": 0.011819940385701125 }, "harness|winogrande|5": { "acc": 0.7348066298342542, "acc_stderr": 0.01240654946619286 } }
配置详情
-
harness_arc_challenge_25
- 分片:2023_08_25T20_49_31.940231
- 路径:
**/details_harness|arc:challenge|25_2023-08-25T20:49:31.940231.parquet
- 路径:
- 分片:2023_09_22T09_17_00.712298
- 路径:
**/details_harness|arc:challenge|25_2023-09-22T09-17-00.712298.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|arc:challenge|25_2023-09-22T09-17-00.712298.parquet
- 路径:
- 分片:2023_08_25T20_49_31.940231
-
harness_drop_3
- 分片:2023_10_15T18_23_36.599949
- 路径:
**/details_harness|drop|3_2023-10-15T18-23-36.599949.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|drop|3_2023-10-15T18-23-36.599949.parquet
- 路径:
- 分片:2023_10_15T18_23_36.599949
-
harness_gsm8k_5
- 分片:2023_10_15T18_23_36.599949
- 路径:
**/details_harness|gsm8k|5_2023-10-15T18-23-36.599949.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|gsm8k|5_2023-10-15T18-23-36.599949.parquet
- 路径:
- 分片:2023_10_15T18_23_36.599949
-
harness_hellaswag_10
- 分片:2023_08_25T20_49_31.940231
- 路径:
**/details_harness|hellaswag|10_2023-08-25T20:49:31.940231.parquet
- 路径:
- 分片:2023_09_22T09_17_00.712298
- 路径:
**/details_harness|hellaswag|10_2023-09-22T09-17-00.712298.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|hellaswag|10_2023-09-22T09-17-00.712298.parquet
- 路径:
- 分片:2023_08_25T20_49_31.940231
-
harness_hendrycksTest_5
- 分片:2023_08_25T20_49_31.940231
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-formal_logic|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-global_facts|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-high_school_biology|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-high_school_geography|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-high_school_physics|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-high_school_psychology|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-high_school_statistics|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-high_school_us_history|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-high_school_world_history|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-human_aging|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-human_sexuality|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-international_law|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-jurisprudence|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-logical_fallacies|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-machine_learning|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-management|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-marketing|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-medical_genetics|5_2023-08-25T20:49:31.940231.parquet - 路径:
**/details_harness|hendrycksTest-miscellaneous|5_2023-08-25T20:49:31.940231.parquet - 路径:`**/details_harness|hendrycksTest-moral_disputes|5_2023-08-25T20:
- 路径:
- 分片:2023_08_25T20_49_31.940231



