Revisiting Residual-free Transformers

CatAstro_Piyush · x · 2026-07-14

This post shares a new blog post titled Revisiting Residual-free Transformers!. The author presents a counterfactual experiment demonstrating that a strong optimizer like Muon might actually be capable of training Transformers without a residual stream.

Original post →

More from Research

• Deep Learning Course Adds Mechanistic Interpretability Lecture This Winter — CSProfKGD · 2026-10-26
• INCLUDE-2.0 benchmark evaluates regional knowledge across 91 languages and 86 countries — ABosselut · 2026-10-26
• CoRL paper: flow-matching VLAs can run 1-step Euler inference with no retraining — kaixhin · 2026-10-26
• CoVer at EMNLP 2026: Training Code LLMs to Generate and Self-Verify Simultaneously — algo_diver · 2026-10-26
• Robotic dexterity should be judged on speed, force control and tool use, not copying humans — DJiafei · 2026-10-26
• Turing-complete malware can run in PyTorch's backward pass via a zero-valued loss — retr0jirachi · 2026-10-26

Research channel →