open-llm-leaderboard-old/details_jondurbin__airoboros-l2-13b-3.0
收藏Hugging Face2023-10-24 更新2024-06-22 收录
下载链接：
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_jondurbin__airoboros-l2-13b-3.0
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是在模型[jondurbin/airoboros-l2-13b-3.0](https://huggingface.co/jondurbin/airoboros-l2-13b-3.0)的评估运行中自动创建的。数据集由64个配置组成，每个配置对应一个评估任务。数据集是从2次运行中创建的，每次运行都可以在特定配置的特定分割中找到，分割名称使用运行的时间戳命名。"train"分割始终指向最新的结果。此外，还有一个名为"results"的配置存储了所有运行的聚合结果，并用于计算和显示在[Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)上的聚合指标。
该数据集是在模型[jondurbin/airoboros-l2-13b-3.0](https://huggingface.co/jondurbin/airoboros-l2-13b-3.0)的评估运行中自动创建的。数据集由64个配置组成，每个配置对应一个评估任务。数据集是从2次运行中创建的，每次运行都可以在特定配置的特定分割中找到，分割名称使用运行的时间戳命名。"train"分割始终指向最新的结果。此外，还有一个名为"results"的配置存储了所有运行的聚合结果，并用于计算和显示在[Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)上的聚合指标。
提供机构：
open-llm-leaderboard-old
原始信息汇总
数据集概述
数据集来源
该数据集是在对模型 jondurbin/airoboros-l2-13b-3.0 进行评估运行期间自动创建的，评估运行在 Open LLM Leaderboard 上进行。
数据集组成
数据集由64个配置组成，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作为一个特定的分片找到，分片名称使用运行的时间戳。"train" 分片始终指向最新的结果。
额外配置
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 Open LLM Leaderboard 上的聚合指标。
数据加载示例
加载特定运行的详细信息，可以使用以下代码： python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_jondurbin__airoboros-l2-13b-3.0", "harness_winogrande_5", split="train")
最新结果
以下是 2023-10-24T07:06:03.975558 运行 的最新结果： python { "all": { "em": 0.20868288590604026, "em_stderr": 0.004161580956848853, "f1": 0.26992973993288605, "f1_stderr": 0.004166447885566019, "acc": 0.4255516933315701, "acc_stderr": 0.009918265858821027 }, "harness|drop|3": { "em": 0.20868288590604026, "em_stderr": 0.004161580956848853, "f1": 0.26992973993288605, "f1_stderr": 0.004166447885566019 }, "harness|gsm8k|5": { "acc": 0.08946171341925702, "acc_stderr": 0.007861583049939738 }, "harness|winogrande|5": { "acc": 0.7616416732438832, "acc_stderr": 0.011974948667702314 } }
配置详情
以下是数据集的配置详情：
-
harness_arc_challenge_25
- 分片：2023_10_08T18_14_31.712178
- 路径：
**/details_harness|arc:challenge|25_2023-10-08T18-14-31.712178.parquet
- 路径：
- 分片：latest
- 路径：
**/details_harness|arc:challenge|25_2023-10-08T18-14-31.712178.parquet
- 路径：
- 分片：2023_10_08T18_14_31.712178
-
harness_drop_3
- 分片：2023_10_24T07_06_03.975558
- 路径：
**/details_harness|drop|3_2023-10-24T07-06-03.975558.parquet
- 路径：
- 分片：latest
- 路径：
**/details_harness|drop|3_2023-10-24T07-06-03.975558.parquet
- 路径：
- 分片：2023_10_24T07_06_03.975558
-
harness_gsm8k_5
- 分片：2023_10_24T07_06_03.975558
- 路径：
**/details_harness|gsm8k|5_2023-10-24T07-06-03.975558.parquet
- 路径：
- 分片：latest
- 路径：
**/details_harness|gsm8k|5_2023-10-24T07-06-03.975558.parquet
- 路径：
- 分片：2023_10_24T07_06_03.975558
-
harness_hellaswag_10
- 分片：2023_10_08T18_14_31.712178
- 路径：
**/details_harness|hellaswag|10_2023-10-08T18-14-31.712178.parquet
- 路径：
- 分片：latest
- 路径：
**/details_harness|hellaswag|10_2023-10-08T18-14-31.712178.parquet
- 路径：
- 分片：2023_10_08T18_14_31.712178
-
harness_hendrycksTest_5
- 分片：2023_10_08T18_14_31.712178
- 路径：
**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-anatomy|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-astronomy|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-business_ethics|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-college_biology|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-college_chemistry|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-college_computer_science|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-college_mathematics|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-college_medicine|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-college_physics|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-computer_security|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-econometrics|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-formal_logic|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-global_facts|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-high_school_biology|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-high_school_european_history|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-high_school_geography|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-high_school_physics|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-high_school_psychology|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-high_school_statistics|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-high_school_us_history|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-high_school_world_history|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-human_aging|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-human_sexuality|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-international_law|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-jurisprudence|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-logical_fallacies|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-machine_learning|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-management|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-marketing|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-medical_genetics|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-miscellaneous|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-moral_disputes|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-moral_scenarios|5_2023-10-08T18-14-31.712178.parquet - 路径：
**/details_harness|hendrycksTest-nutrition|5_2023-10-08T18-14-31.712178.parquet - 路径：`**/details_harness|hendrycksTest-philosophy|5_2023-10-08T1
- 路径：
- 分片：2023_10_08T18_14_31.712178
