open-llm-leaderboard-old/details_yihan6324__llama2-7b-instructmining-orca-40k
收藏数据集概述
该数据集是在对模型 yihan6324/llama2-7b-instructmining-orca-40k 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 61 个配置,每个配置对应一个评估任务。
- 数据集由 1 次运行创建,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 "results" 配置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_yihan6324__llama2-7b-instructmining-orca-40k", "harness_truthfulqa_mc_0", split="train")
最新结果
这些是最新结果(来自 2023-08-18T00:53:27.654117 运行)的示例: python { "all": { "acc": 0.4847120233306423, "acc_stderr": 0.03527399847085323, "acc_norm": 0.4884455010512822, "acc_norm_stderr": 0.035257414280301984, "mc1": 0.36107711138310894, "mc1_stderr": 0.016814312844836882, "mc2": 0.5103220670450638, "mc2_stderr": 0.015890639542177364 }, "harness|arc:challenge|25": { "acc": 0.5298634812286689, "acc_stderr": 0.014585305840007105, "acc_norm": 0.5674061433447098, "acc_norm_stderr": 0.014478005694182524 }, "harness|hellaswag|10": { "acc": 0.6196972714598685, "acc_stderr": 0.004844690404713595, "acc_norm": 0.8024297948615814, "acc_norm_stderr": 0.0039735233080143454 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.35, "acc_stderr": 0.0479372485441102, "acc_norm": 0.35, "acc_norm_stderr": 0.0479372485441102 }, # 其他任务的结果... }
配置详情
-
harness_arc_challenge_25
- 分割:2023_08_18T00_53_27.654117
- 路径:
**/details_harness|arc:challenge|25_2023-08-18T00:53:27.654117.parquet - 分割:latest
- 路径:
**/details_harness|arc:challenge|25_2023-08-18T00:53:27.654117.parquet
-
harness_hellaswag_10
- 分割:2023_08_18T00_53_27.654117
- 路径:
**/details_harness|hellaswag|10_2023-08-18T00:53:27.654117.parquet - 分割:latest
- 路径:
**/details_harness|hellaswag|10_2023-08-18T00:53:27.654117.parquet
-
harness_hendrycksTest_5
- 分割:2023_08_18T00_53_27.654117
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-anatomy|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-astronomy|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-business_ethics|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-college_biology|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-college_chemistry|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-college_computer_science|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-college_mathematics|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-college_medicine|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-college_physics|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-computer_security|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-econometrics|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-formal_logic|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-global_facts|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-high_school_biology|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-18T00:53:27.654117.parquet**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-18T00:53:27.654117.parquet
其他路径...
}



