Hugging Face Papers ↗ · 2026-11-13
WeightBench:评估语言模型作为权重空间程序员的能力
Media reports Hugging Face Papers Source language: Chinese
论文摘要(作者报告,未经本站复测):近期工作表明大语言模型可以直接操作其他模型的参数。
Full text
The WEIGHTBENCH benchmark evaluates the capacity of large language models to perform Weight-Space Programming tasks, which involve reading, modifying, and generating the weights of various neural network architectures.