open-llm-leaderboard/details_CalderaAI__30B-Lazarus
收藏数据集概述
该数据集是在对模型 CalderaAI/30B-Lazarus 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 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" 分割始终指向最新的结果。
- 额外的配置 "results" 存储所有运行结果的聚合,用于计算和显示 Open LLM Leaderboard 上的聚合指标。
加载数据集示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_CalderaAI__30B-Lazarus", "harness_winogrande_5", split="train")
最新结果
以下是 2023-10-13T10:30:29.206402 运行 的最新结果:
python { "all": { "em": 0.15866191275167785, "em_stderr": 0.0037416337044887996, "f1": 0.2289985318791943, "f1_stderr": 0.003861278919536814, "acc": 0.43053621617869603, "acc_stderr": 0.009464164192315844 }, "harness|drop|3": { "em": 0.15866191275167785, "em_stderr": 0.0037416337044887996, "f1": 0.2289985318791943, "f1_stderr": 0.003861278919536814 }, "harness|gsm8k|5": { "acc": 0.07733131159969674, "acc_stderr": 0.007357713523222348 }, "harness|winogrande|5": { "acc": 0.7837411207576953, "acc_stderr": 0.01157061486140934 } }
配置详情
配置列表
harness_arc_challenge_25harness_drop_3harness_gsm8k_5harness_hellaswag_10harness_hendrycksTest_5harness_hendrycksTest_abstract_algebra_5harness_hendrycksTest_anatomy_5harness_hendrycksTest_astronomy_5harness_hendrycksTest_business_ethics_5harness_hendrycksTest_clinical_knowledge_5harness_hendrycksTest_college_biology_5harness_hendrycksTest_college_chemistry_5harness_hendrycksTest_college_computer_science_5harness_hendrycksTest_college_mathematics_5harness_hendrycksTest_college_medicine_5harness_hendrycksTest_college_physics_5harness_hendrycksTest_computer_security_5harness_hendrycksTest_conceptual_physics_5harness_hendrycksTest_econometrics_5harness_hendrycksTest_electrical_engineering_5harness_hendrycksTest_elementary_mathematics_5
数据文件路径
每个配置包含多个分割,每个分割对应不同的运行时间戳。例如:
-
harness_arc_challenge_25配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_drop_3配置包含以下分割:2023_10_13T10_30_29.206402latest
-
harness_gsm8k_5配置包含以下分割:2023_10_13T10_30_29.206402latest
-
harness_hellaswag_10配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_abstract_algebra_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_anatomy_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_astronomy_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_business_ethics_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_clinical_knowledge_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_college_biology_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_college_chemistry_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_college_computer_science_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_college_mathematics_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_college_medicine_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_college_physics_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_computer_security_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_conceptual_physics_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_econometrics_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_electrical_engineering_5配置包含以下分割:2023_07_19T22_16_39.327210latest
-
harness_hendrycksTest_elementary_mathematics_5配置包含以下分割:2023_07_19T22_16_39.327210latest



