open-llm-leaderboard-old/details_EleutherAI__pythia-70m-deduped
收藏数据集概述
数据集简介
- 名称: Evaluation run of EleutherAI/pythia-70m-deduped
- 来源: 自动创建于模型 EleutherAI/pythia-70m-deduped 在 Open LLM Leaderboard 的评估运行期间。
数据集结构
- 配置数量: 64
- 配置详情: 每个配置对应一个评估任务。
- 数据来源: 从2次运行中创建。每个运行在每个配置中作为一个特定的分片存在,分片名称使用运行的时间戳。"train" 分片始终指向最新的结果。
数据集加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_EleutherAI__pythia-70m-deduped", "harness_winogrande_5", split="train")
最新结果
- 来源: 最新结果来自2023-10-19T00:18:19.073831
- 结果示例: python { "all": { "em": 0.0012583892617449664, "em_stderr": 0.0003630560893119184, "f1": 0.023000209731543642, "f1_stderr": 0.0009427318515971101, "acc": 0.24822415153906865, "acc_stderr": 0.007026065573457934 }, "harness|drop|3": { "em": 0.0012583892617449664, "em_stderr": 0.0003630560893119184, "f1": 0.023000209731543642, "f1_stderr": 0.0009427318515971101 }, "harness|gsm8k|5": { "acc": 0.0, "acc_stderr": 0.0 }, "harness|winogrande|5": { "acc": 0.4964483030781373, "acc_stderr": 0.014052131146915867 } }
配置详情
-
配置名称: harness_arc_challenge_25
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|arc:challenge|25_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_drop_3
- 分片: 2023_10_19T00_18_19.073831, latest
- 路径: **/details_harness|drop|3_2023-10-19T00-18-19.073831.parquet
-
配置名称: harness_gsm8k_5
- 分片: 2023_10_19T00_18_19.073831, latest
- 路径: **/details_harness|gsm8k|5_2023-10-19T00-18-19.073831.parquet
-
配置名称: harness_hellaswag_10
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hellaswag|10_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: 包含多个学科的详细路径
-
配置名称: harness_hendrycksTest_abstract_algebra_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_anatomy_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-anatomy|5_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_astronomy_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-astronomy|5_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_business_ethics_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-business_ethics|5_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_clinical_knowledge_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_college_biology_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-college_biology|5_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_college_chemistry_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_college_computer_science_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_college_mathematics_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_college_medicine_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-college_medicine|5_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_college_physics_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-college_physics|5_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_computer_security_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-computer_security|5_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_conceptual_physics_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_econometrics_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-econometrics|5_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_electrical_engineering_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T13:42:51.890470.parquet
-
配置名称: harness_hendrycksTest_elementary_mathematics_5
- 分片: 2023_07_19T13_42_51.890470, latest
- 路径: **/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T13:42:51.890470.parquet
以上是数据集的详细概述,包括数据集的来源、结构、加载示例、最新结果以及各个配置的详细信息。



