open-llm-leaderboard/details_ehartford__dolphin-llama-13b
收藏数据集概述
数据集名称
Evaluation run of ehartford/dolphin-llama-13b
数据集摘要
该数据集是在评估模型 ehartford/dolphin-llama-13b 在 Open LLM Leaderboard 上的自动创建的。
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4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_ehartford__dolphin-llama-13b", "harness_winogrande_5", split="train")
最新结果
这些是最新结果(来自2023-10-22T08:31:06.423580运行)的示例: python { "all": { "em": 0.09867869127516779, "em_stderr": 0.003054155613095951, "f1": 0.1882760067114087, "f1_stderr": 0.0033481950499125467, "acc": 0.42166909111145284, "acc_stderr": 0.011280060733885005 }, "harness|drop|3": { "em": 0.09867869127516779, "em_stderr": 0.003054155613095951, "f1": 0.1882760067114087, "f1_stderr": 0.0033481950499125467 }, "harness|gsm8k|5": { "acc": 0.14404852160727824, "acc_stderr": 0.009672110973065284 }, "harness|winogrande|5": { "acc": 0.6992896606156275, "acc_stderr": 0.012888010494704725 } }
配置详情
-
harness_arc_challenge_25
- 分片:2023_07_24T12_30_40.142317
- 路径:
**/details_harness|arc:challenge|25_2023-07-24T12:30:40.142317.parquet
- 路径:
- 分片:2023_07_24T16_19_11.269492
- 路径:
**/details_harness|arc:challenge|25_2023-07-24T16:19:11.269492.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|arc:challenge|25_2023-07-24T16:19:11.269492.parquet
- 路径:
- 分片:2023_07_24T12_30_40.142317
-
harness_drop_3
- 分片:2023_10_22T01_25_54.857200
- 路径:
**/details_harness|drop|3_2023-10-22T01-25-54.857200.parquet
- 路径:
- 分片:2023_10_22T08_31_06.423580
- 路径:
**/details_harness|drop|3_2023-10-22T08-31-06.423580.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|drop|3_2023-10-22T08-31-06.423580.parquet
- 路径:
- 分片:2023_10_22T01_25_54.857200
-
harness_gsm8k_5
- 分片:2023_10_22T01_25_54.857200
- 路径:
**/details_harness|gsm8k|5_2023-10-22T01-25-54.857200.parquet
- 路径:
- 分片:2023_10_22T08_31_06.423580
- 路径:
**/details_harness|gsm8k|5_2023-10-22T08-31-06.423580.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|gsm8k|5_2023-10-22T08-31-06.423580.parquet
- 路径:
- 分片:2023_10_22T01_25_54.857200
-
harness_hellaswag_10
- 分片:2023_07_24T12_30_40.142317
- 路径:
**/details_harness|hellaswag|10_2023-07-24T12:30:40.142317.parquet
- 路径:
- 分片:2023_07_24T16_19_11.269492
- 路径:
**/details_harness|hellaswag|10_2023-07-24T16:19:11.269492.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|hellaswag|10_2023-07-24T16:19:11.269492.parquet
- 路径:
- 分片:2023_07_24T12_30_40.142317
-
harness_hendrycksTest_5
- 分片:2023_07_24T12_30_40.142317
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-formal_logic|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-global_facts|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-high_school_biology|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-high_school_european_history|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-high_school_geography|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-high_school_physics|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-high_school_psychology|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-high_school_statistics|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-high_school_us_history|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-high_school_world_history|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-human_aging|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-human_sexuality|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-international_law|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-jurisprudence|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-logical_fallacies|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-machine_learning|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-management|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-marketing|5_2023-07-24T12:30:40.142317.parquet - 路径:
**/details_harness|hendrycksTest-medical_genetics|5_2023-07-24T12:30:40.142317.parquet - 路径:`**/details_har
- 路径:
- 分片:2023_07_24T12_30_40.142317



