piyush314/ghosts-of-softmax
Reproducibility code for: Ghosts of Softmax
4 best points · 3 comments · 1 Hacker News post · 3 GitHub stars · View on GitHub
Hacker News discussions
- Ghosts of Softmax: Zeros of the partition function explain training instability (4 points, 3 comments, by g4omingron, 2026-03-17)