open-llm-leaderboard/details_Writer__InstructPalmyra-20b
收藏数据集概述
数据集名称
- Evaluation run of Writer/InstructPalmyra-20b
数据集描述
- 该数据集是在模型 Writer/InstructPalmyra-20b 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,每次运行可以在每个配置中找到一个特定的分割,分割名称使用运行的时间戳。
- "train" 分割始终指向最新的结果。
- 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_Writer__InstructPalmyra-20b", "harness_winogrande_5", split="train")
最新结果
- 这些是最新结果,来自 2023-10-12T18:44:54.114721 的运行: python { "all": { "em": 0.01929530201342282, "em_stderr": 0.0014087520774405866, "f1": 0.08355075503355748, "f1_stderr": 0.0019786281704718585, "acc": 0.33648760481410367, "acc_stderr": 0.008897385527705382 }, "harness|drop|3": { "em": 0.01929530201342282, "em_stderr": 0.0014087520774405866, "f1": 0.08355075503355748, "f1_stderr": 0.0019786281704718585 }, "harness|gsm8k|5": { "acc": 0.02577710386656558, "acc_stderr": 0.004365042953621808 }, "harness|winogrande|5": { "acc": 0.6471981057616417, "acc_stderr": 0.013429728101788956 } }
配置详情
-
harness_arc_challenge_25
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|arc:challenge|25_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|arc:challenge|25_2023-08-29T16:04:46.105936.parquet
-
harness_drop_3
- 分割:2023_10_12T18_44_54.114721
- 路径:
**/details_harness|drop|3_2023-10-12T18-44-54.114721.parquet - 分割:latest
- 路径:
**/details_harness|drop|3_2023-10-12T18-44-54.114721.parquet
-
harness_gsm8k_5
- 分割:2023_10_12T18_44_54.114721
- 路径:
**/details_harness|gsm8k|5_2023-10-12T18-44-54.114721.parquet - 分割:latest
- 路径:
**/details_harness|gsm8k|5_2023-10-12T18-44-54.114721.parquet
-
harness_hellaswag_10
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hellaswag|10_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hellaswag|10_2023-08-29T16:04:46.105936.parquet
-
harness_hendrycksTest_5
- 分割:2023_08_29T16_04_46.105936
- 路径:多个路径,如
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-29T16:04:46.105936.parquet等 - 分割:latest
- 路径:多个路径,如
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-29T16:04:46.105936.parquet等
-
harness_hendrycksTest_abstract_algebra_5
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-29T16:04:46.105936.parquet
-
harness_hendrycksTest_anatomy_5
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-08-29T16:04:46.105936.parquet
-
harness_hendrycksTest_astronomy_5
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-08-29T16:04:46.105936.parquet
-
harness_hendrycksTest_business_ethics_5
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-08-29T16:04:46.105936.parquet
-
harness_hendrycksTest_clinical_knowledge_5
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-29T16:04:46.105936.parquet
-
harness_hendrycksTest_college_biology_5
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-08-29T16:04:46.105936.parquet
-
harness_hendrycksTest_college_chemistry_5
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-08-29T16:04:46.105936.parquet
-
harness_hendrycksTest_college_computer_science_5
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-08-29T16:04:46.105936.parquet
-
harness_hendrycksTest_college_mathematics_5
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-08-29T16:04:46.105936.parquet
-
harness_hendrycksTest_college_medicine_5
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-08-29T16:04:46.105936.parquet
-
harness_hendrycksTest_college_physics_5
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-08-29T16:04:46.105936.parquet
-
harness_hendrycksTest_computer_security_5
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-08-29T16:04:46.105936.parquet
-
harness_hendrycksTest_conceptual_physics_5
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-29T16:04:46.105936.parquet
-
harness_hendrycksTest_econometrics_5
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-08-29T16:04:46.105936.parquet
-
harness_hendrycksTest_electrical_engineering_5
- 分割:2023_08_29T16_04_46.105936
- 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-29T16:04:46.105936.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-29T16:04:46.105936.parquet
-
**har



