Noah A. Smith

Latest

• Saturated Transformers are Constant-Depth Threshold Circuits
• Competency Problems: On Finding and Removing Artifacts in Language Data
• Effects of Parameter Norm Growth During Transformer Training: Inductive Bias from Gradient Descent
• Provable Limitations of Acquiring Meaning from Ungrounded Form: What Will Future Language Models Understand?
• A Formal Hierarchy of RNN Architectures