open-llm-leaderboard-old/details_vibhorag101__llama-2-13b-chat-hf-phr_mental_therapy
收藏数据集概述
该数据集是在对模型 vibhorag101/llama-2-13b-chat-hf-phr_mental_therapy 进行评估运行期间自动创建的。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,每个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。
数据集结构
配置
- results: 存储所有运行结果的聚合结果,用于计算和显示在 Open LLM Leaderboard 上的聚合指标。
加载数据
以下是加载特定运行详细信息的示例代码: python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_vibhorag101__llama-2-13b-chat-hf-phr_mental_therapy", "harness_winogrande_5", split="train")
最新结果
以下是 2023-12-04T18:26:43.065214 运行 的最新结果:
python { "all": { "acc": 0.2434367192235923, "acc_stderr": 0.03008501938303984, "acc_norm": 0.24224538912156782, "acc_norm_stderr": 0.030747150403453674, "mc1": 0.2778457772337821, "mc1_stderr": 0.015680929364024643, "mc2": 0.4692403294958895, "mc2_stderr": 0.015061938982346217 }, "harness|arc:challenge|25": { "acc": 0.36945392491467577, "acc_stderr": 0.014104578366491894, "acc_norm": 0.38822525597269625, "acc_norm_stderr": 0.01424161420741405 }, "harness|hellaswag|10": { "acc": 0.5696076478789086, "acc_stderr": 0.004941191607317913, "acc_norm": 0.7276438956383191, "acc_norm_stderr": 0.004442623590846322 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.22, "acc_stderr": 0.04163331998932268, "acc_norm": 0.22, "acc_norm_stderr": 0.04163331998932268 }, # 其他任务的结果... }
配置详情
配置列表
- harness_arc_challenge_25
- harness_gsm8k_5
- harness_hellaswag_10
- harness_hendrycksTest_5
数据文件路径
每个配置包含多个数据文件路径,例如:
-
harness_arc_challenge_25:
**/details_harness|arc:challenge|25_2023-12-04T18-26-43.065214.parquet**/details_harness|arc:challenge|25_2023-12-04T18-26-43.065214.parquet
-
harness_gsm8k_5:
**/details_harness|gsm8k|5_2023-12-04T18-26-43.065214.parquet**/details_harness|gsm8k|5_2023-12-04T18-26-43.065214.parquet
-
harness_hellaswag_10:
**/details_harness|hellaswag|10_2023-12-04T18-26-43.065214.parquet**/details_harness|hellaswag|10_2023-12-04T18-26-43.065214.parquet
-
harness_hendrycksTest_5:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-anatomy|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-astronomy|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-business_ethics|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-college_biology|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-college_chemistry|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-college_computer_science|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-college_mathematics|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-college_medicine|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-college_physics|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-computer_security|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-econometrics|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-formal_logic|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-global_facts|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-high_school_biology|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-high_school_chemistry|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-high_school_computer_science|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-high_school_european_history|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-high_school_geography|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-high_school_mathematics|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-high_school_physics|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-high_school_psychology|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-high_school_statistics|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-high_school_us_history|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-high_school_world_history|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-human_aging|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-human_sexuality|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-international_law|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-jurisprudence|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-logical_fallacies|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-machine_learning|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-management|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-marketing|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-medical_genetics|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-miscellaneous|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-moral_disputes|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-moral_scenarios|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-nutrition|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-philosophy|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-prehistory|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-professional_accounting|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-professional_law|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-professional_medicine|5_2023-12-04T18-26-43.065214.parquet**/details_harness|hendrycksTest-professional_psychology|5_2023-12-04T18-26-43.065214.parquet- `**/details_harness|



