open-llm-leaderboard-old/details_frankenmerger__gemoy-4b-instruct-scientific
收藏数据集概述
数据集摘要
该数据集是在评估模型 frankenmerger/gemoy-4b-instruct-scientific 在 Open LLM Leaderboard 上的运行过程中自动创建的。
数据集组成
- 该数据集包含 63 个配置,每个配置对应一个评估任务。
- 数据集从 1 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_frankenmerger__gemoy-4b-instruct-scientific", "harness_winogrande_5", split="train")
最新结果
以下是 2024-03-10T11:10:43.531199 运行的最新结果:
python { "all": { "acc": 0.3887480937200939, "acc_stderr": 0.033967527013847434, "acc_norm": 0.3919353670094879, "acc_norm_stderr": 0.03472007289325813, "mc1": 0.26438188494492043, "mc1_stderr": 0.015438211119522514, "mc2": 0.4195962328166831, "mc2_stderr": 0.014414337460874078 }, "harness|arc:challenge|25": { "acc": 0.39334470989761094, "acc_stderr": 0.014275101465693026, "acc_norm": 0.4197952218430034, "acc_norm_stderr": 0.014422181226303026 }, "harness|hellaswag|10": { "acc": 0.46106353316072496, "acc_stderr": 0.004974628903829138, "acc_norm": 0.6304521011750648, "acc_norm_stderr": 0.004816958817726085 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.33, "acc_stderr": 0.047258156262526045, "acc_norm": 0.33, "acc_norm_stderr": 0.047258156262526045 }, # 其他任务的结果... }
配置详情
-
config_name: harness_arc_challenge_25
- 数据文件:
- 分割: 2024_03_10T11_10_43.531199
- 路径:
**/details_harness|arc:challenge|25_2024-03-10T11-10-43.531199.parquet
- 路径:
- 分割: latest
- 路径:
**/details_harness|arc:challenge|25_2024-03-10T11-10-43.531199.parquet
- 路径:
- 分割: 2024_03_10T11_10_43.531199
- 数据文件:
-
config_name: harness_gsm8k_5
- 数据文件:
- 分割: 2024_03_10T11_10_43.531199
- 路径:
**/details_harness|gsm8k|5_2024-03-10T11-10-43.531199.parquet
- 路径:
- 分割: latest
- 路径:
**/details_harness|gsm8k|5_2024-03-10T11-10-43.531199.parquet
- 路径:
- 分割: 2024_03_10T11_10_43.531199
- 数据文件:
-
config_name: harness_hellaswag_10
- 数据文件:
- 分割: 2024_03_10T11_10_43.531199
- 路径:
**/details_harness|hellaswag|10_2024-03-10T11-10-43.531199.parquet
- 路径:
- 分割: latest
- 路径:
**/details_harness|hellaswag|10_2024-03-10T11-10-43.531199.parquet
- 路径:
- 分割: 2024_03_10T11_10_43.531199
- 数据文件:
-
config_name: harness_hendrycksTest_5
- 数据文件:
- 分割: 2024_03_10T11_10_43.531199
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-anatomy|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-astronomy|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-business_ethics|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-college_biology|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-college_chemistry|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-college_computer_science|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-college_mathematics|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-college_medicine|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-college_physics|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-computer_security|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-econometrics|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2024-03-10T11-10-43.531199.parquet**/details_harness|hendrycksTest-formal_logic|5_2024-03-10T11-10-43.531199.parquet
- 路径:
- 分割: 2024_03_10T11_10_43.531199
- 数据文件:



