安裝中文字典英文字典辭典工具!
安裝中文字典英文字典辭典工具!
|
- cais mmlu · Datasets at Hugging Face
Dataset Card for MMLU Dataset Summary Measuring Massive Multitask Language Understanding by Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, and Jacob Steinhardt (ICLR 2021) This is a massive multitask test consisting of multiple-choice questions from various branches of knowledge
- AIGC每周精选--大模型评测之MMLU-PRO与MMLU - 知乎
MMLU （Massive Multitask Language Understanding）是一个新的基准，用于衡量在 零样本 （zero-shot）和 少样本 （few-shot）情形下，大模型在预训练期间获得的世界知识。
- MMLU - Wikipedia
Measuring Massive Multitask Language Understanding (MMLU) is a popular benchmark for evaluating the capabilities of large language models It inspired several other versions and spin-offs, such as MMLU-Pro, MMMLU and MMLU-Redux
- GitHub - standardgalactic mmlu: Measuring Massive Multitask Language . . .
Measuring Massive Multitask Language Understanding | ICLR 2021 - standardgalactic mmlu
- LLMs：MMLU基准 (大规模多任务语言理解)的简介 (MMLU-ZS MMLU-FS)、安装、使用方法之详细攻略
MMLU是一个多任务语言理解基准，通过MMLU-ZS和MMLU-FS测试设置评估模型的泛化和知识迁移能力。 文章介绍了如何使用预训练模型进行评估，并指出该基准有助于揭示模型在学术和专业领域的不足。
- 大模型MMLU评测基准详情以及最新排行结果 | 数据学习 (DataLearner)
一个涵盖 57 个主题的多项选择题基准，用于评估大规模语言模型的知识和推理能力。 。 查看MMLU介绍、评测指标、官方数据集链接、详细测试结果及大模型排名，掌握 AI 评测趋势！
- MMLU_百度百科
MMLU（Massive Multitask Language Understanding）是由加州大学伯克利分校研究人员于2020年9月推出的大模型测评基准，包含初等数学、计算机科学、法律、历史等57项学科任务，旨在评估大模型的跨领域知识覆盖和复杂问题解决能力。
- [2009. 03300] Measuring Massive Multitask Language Understanding - arXiv. org
We propose a new test to measure a text model's multitask accuracy The test covers 57 tasks including elementary mathematics, US history, computer science, law, and more To attain high accuracy on this test, models must possess extensive world knowledge and problem solving ability We find that while most recent models have near random-chance accuracy, the very largest GPT-3 model improves
|
|
|