open-llm-leaderboard/details_luffycodes__nash-vicuna-13b-v1dot5-ep2-w-rag-w-simple

该数据集是在评估luffycodes/nash-vicuna-13b-v1dot5-ep2-w-rag-w-simple模型运行期间自动创建的，包含64个配置，每个配置对应一个评估任务。数据集由2次运行创建，每次运行作为一个特定的时间戳命名的分割。train分割始终指向最新结果。额外的results配置存储了运行中的聚合结果，用于在Open LLM Leaderboard上计算和显示聚合指标。

原始信息汇总

数据集概述

该数据集是在对模型 luffycodes/nash-vicuna-13b-v1dot5-ep2-w-rag-w-simple 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。每个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。
• 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

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_luffycodes__nash-vicuna-13b-v1dot5-ep2-w-rag-w-simple", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-15T01:13:17.966803 运行的最新结果： python { "all": { "em": 0.18760486577181207, "em_stderr": 0.003998023634854269, "f1": 0.2689041526845642, "f1_stderr": 0.00405255679434132, "acc": 0.4260142426906131, "acc_stderr": 0.010340665159137691 }, "harness|drop|3": { "em": 0.18760486577181207, "em_stderr": 0.003998023634854269, "f1": 0.2689041526845642, "f1_stderr": 0.00405255679434132 }, "harness|gsm8k|5": { "acc": 0.10538286580742987, "acc_stderr": 0.008457575884041755 }, "harness|winogrande|5": { "acc": 0.7466456195737964, "acc_stderr": 0.012223754434233626 } }

配置详情

• harness_arc_challenge_25

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|arc:challenge|25_2023-09-01T18:20:29.445308.parquet

• harness_drop_3

  • 分片: 2023_10_15T01_13_17.966803, latest
  • 路径: **/details_harness|drop|3_2023-10-15T01-13-17.966803.parquet

• harness_gsm8k_5

  • 分片: 2023_10_15T01_13_17.966803, latest
  • 路径: **/details_harness|gsm8k|5_2023-10-15T01-13-17.966803.parquet

• harness_hellaswag_10

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hellaswag|10_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: 多个路径，例如 **/details_harness|hendrycksTest-abstract_algebra|5_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_abstract_algebra_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hendrycksTest-abstract_algebra|5_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_anatomy_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hendrycksTest-anatomy|5_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_astronomy_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hendrycksTest-astronomy|5_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_business_ethics_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hendrycksTest-business_ethics|5_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_clinical_knowledge_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_college_biology_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hendrycksTest-college_biology|5_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_college_chemistry_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hendrycksTest-college_chemistry|5_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_college_computer_science_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hendrycksTest-college_computer_science|5_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_college_mathematics_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hendrycksTest-college_mathematics|5_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_college_medicine_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hendrycksTest-college_medicine|5_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_college_physics_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hendrycksTest-college_physics|5_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_computer_security_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hendrycksTest-computer_security|5_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_conceptual_physics_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hendrycksTest-conceptual_physics|5_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_econometrics_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hendrycksTest-econometrics|5_2023-09-01T18:20:29.445308.parquet

• harness_hendrycksTest_electrical_engineering_5

  • 分片: 2023_09_01T18_20_29.445308, latest
  • 路径: **/details_harness|hendrycksTest-electrical_engineering|5_2023-09-01T18:20:29.445308.parquet

以上是数据集的详细概述，包括数据集的组成、加载示例、最新结果以及各个配置的详细信息。