open-llm-leaderboard-old/details_togethercomputer__GPT-NeoXT-Chat-Base-20B
收藏数据集概述
数据集简介
该数据集是在评估模型 togethercomputer/GPT-NeoXT-Chat-Base-20B 在 Open LLM Leaderboard 上的自动创建的。
数据集结构
- 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特定的分割,分割名称使用运行的时间戳。
- "train" 分割总是指向最新的结果。
- 一个额外的配置 "results" 存储所有运行的聚合结果,用于计算和显示聚合指标在 Open LLM Leaderboard 上。
数据加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_togethercomputer__GPT-NeoXT-Chat-Base-20B", "harness_winogrande_5", split="train")
最新结果
以下是 最新结果从 run 2023-10-21T19:34:18.998946 的数据: python { "all": { "em": 0.0014681208053691276, "em_stderr": 0.0003921042190298267, "f1": 0.0505337667785235, "f1_stderr": 0.001244239988388026, "acc": 0.3699338727947376, "acc_stderr": 0.010093691461965018 }, "harness|drop|3": { "em": 0.0014681208053691276, "em_stderr": 0.0003921042190298267, "f1": 0.0505337667785235, "f1_stderr": 0.001244239988388026 }, "harness|gsm8k|5": { "acc": 0.06899166034874905, "acc_stderr": 0.00698099583483858 }, "harness|winogrande|5": { "acc": 0.6708760852407262, "acc_stderr": 0.013206387089091458 } }
配置详情
以下是数据集的配置详情:
-
harness_arc_challenge_25
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|arc:challenge|25_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|arc:challenge|25_2023-07-19T21:40:44.259947.parquet
-
harness_drop_3
- 分割:2023_10_21T19_34_18.998946
- 路径:
**/details_harness|drop|3_2023-10-21T19-34-18.998946.parquet - 分割:latest
- 路径:
**/details_harness|drop|3_2023-10-21T19-34-18.998946.parquet
-
harness_gsm8k_5
- 分割:2023_10_21T19_34_18.998946
- 路径:
**/details_harness|gsm8k|5_2023-10-21T19-34-18.998946.parquet - 分割:latest
- 路径:
**/details_harness|gsm8k|5_2023-10-21T19-34-18.998946.parquet
-
harness_hellaswag_10
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hellaswag|10_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|hellaswag|10_2023-07-19T21:40:44.259947.parquet
-
harness_hendrycksTest_5
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T21:40:44.259947.parquet等 40 个文件 - 分割:latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T21:40:44.259947.parquet等 40 个文件
-
harness_hendrycksTest_abstract_algebra_5
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T21:40:44.259947.parquet
-
harness_hendrycksTest_anatomy_5
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-07-19T21:40:44.259947.parquet
-
harness_hendrycksTest_astronomy_5
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-07-19T21:40:44.259947.parquet
-
harness_hendrycksTest_business_ethics_5
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T21:40:44.259947.parquet
-
harness_hendrycksTest_clinical_knowledge_5
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T21:40:44.259947.parquet
-
harness_hendrycksTest_college_biology_5
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-07-19T21:40:44.259947.parquet
-
harness_hendrycksTest_college_chemistry_5
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T21:40:44.259947.parquet
-
harness_hendrycksTest_college_computer_science_5
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T21:40:44.259947.parquet
-
harness_hendrycksTest_college_mathematics_5
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T21:40:44.259947.parquet
-
harness_hendrycksTest_college_medicine_5
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T21:40:44.259947.parquet
-
harness_hendrycksTest_college_physics_5
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-07-19T21:40:44.259947.parquet
-
harness_hendrycksTest_computer_security_5
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-07-19T21:40:44.259947.parquet
-
harness_hendrycksTest_conceptual_physics_5
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T21:40:44.259947.parquet
-
harness_hendrycksTest_econometrics_5
- 分割:2023_07_19T21_40_44.259947
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-07-19T21:40:44.259947.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-07-19T21:40:44.259947.parquet
-
harness_hendrycksTest_electrical_engineering_5
- 分割:2023_07_19T21_40_44.259947
- 路径:`**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T21:40:44.259947.par



