Jon Barron: every softmax & cross-entropy implementation (JAX, PyTorch) has two subtle bugs
jon_barron · x · 2026-10-25
Google researcher Jon Barron shares a new result: all existing implementations of softmax and cross-entropy — in JAX, PyTorch, textbooks, etc. — contain two subtle bugs. A simple rewrite fixes both and eliminates 'slingshots' during grokking.
@ritwikak reports the fix just prevented a training collapse in her own model run, calling the implications profound.
A numerical issue embedded in every major framework, affecting countless training runs, has now been pinpointed with a concrete fix — directly relevant to training-stability research.
More from Research
- Paper: "bottleneck pressure" metric shows where AI-infeasible work gains value as AI speeds up the rest — soumitrashukla9 · 2026-10-28
- New IEEE IT paper: geometric Jensen-Shannon divergence between Gaussian measures on Hilbert space — FrnkNlsn · 2026-10-28
- One-page Q-learning notes with math, code and a playground for RL beginners — East-Muffin-6472 · 2026-10-28
- CPR-Dataset: retrievers fail at cross-domain discovery, training gives +70% on BRIGHT — coallaoh · 2026-10-28
- Hugging Face robot duck walks in simulated snow with 290,000 physics-simulated flakes — lukas_m_ziegler · 2026-10-28
- Mercor, Stanford and NBER launch TaskNet, an open-source task catalog updating O*NET for the AI era — soumitrashukla9 · 2026-10-28