open-llm-leaderboard/details_MaziyarPanahi__Llama-3-8B-Instruct-v0.3

该数据集是在Open LLM Leaderboard上对模型MaziyarPanahi/Llama-3-8B-Instruct-v0.3进行评估时自动创建的。数据集由63个配置组成，每个配置对应一个评估任务。数据集由3次运行创建，每次运行的结果作为特定配置中的一个分割，分割名称使用运行的时间戳。train分割始终指向最新的结果。此外，还有一个名为results的配置，存储了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。

原始信息汇总

数据集概述

数据集名称

• pretty_name: Evaluation run of MaziyarPanahi/Llama-3-8B-Instruct-v0.3

数据集创建

• dataset_summary: Dataset automatically created during the evaluation run of model MaziyarPanahi/Llama-3-8B-Instruct-v0.3 on the Open LLM Leaderboard.

数据集结构

• Composition: Composed of 63 configurations, each corresponding to one of the evaluated tasks.
• Runs: Created from 3 runs. Each run is found as a specific split in each configuration, named using the timestamp of the run.
• Splits: The "train" split always points to the latest results.
• Additional Configuration: "results" stores all the aggregated results of the run, used to compute and display aggregated metrics on the Open LLM Leaderboard.

数据集加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_MaziyarPanahi__Llama-3-8B-Instruct-v0.3", "harness_truthfulqa_mc_0", split="train")

最新结果

• Latest Results: latest results from run 2024-05-07T15:22:57.619707

数据集配置详情

配置列表

• harness_arc_challenge_25
• harness_gsm8k_5
• harness_hellaswag_10
• harness_hendrycksTest_5

每个配置包含多个数据文件，每个文件对应不同的运行时间戳，确保数据的最新性和完整性。