Training Vision Models for LLM-Driven Weight-Space Programming Paper • 2611.26408 • Published 4 days ago • 17
Approximating Softmax in Pretrained LLMs: Model Sensitivity and Kernel Acceleration Paper • 2609.33586 • Published Sep 27 • 4
Pretraining Transformers with Quantized Softmax in Attention Paper • 2609.33591 • Published Sep 27 • 7