Detecting model weight tampering in GPU memory (Rome attacks)
rafaeldavidtin
18 points
1 comment
October 19, 2026
Related Discussions
Found 5 related stories in 107.9ms across 9,963 title embeddings via pgvector HNSW
- What happens when a GPU reads memory? somnial · 15 pts · August 13, 2026 · 52% similar
- What happens when a GPU reads memory ibobev · 105 pts · August 21, 2026 · 52% similar
- Lossless model compression experiment: GLM-5.2 in 25% less memory hambandit · 16 pts · July 20, 2026 · 50% similar
- Every Model Cheats vga805 · 91 pts · August 20, 2026 · 50% similar
- Benchmarking 15 “E-Waste” GPUs with Modern Workloads eso_logic · 123 pts · July 13, 2026 · 49% similar
Discussion Highlights (1 comments)
altruios
Curious. This feels like a tool to add to an LLM, being able to modify it's own weights, to be able to relearn, refine its own understanding.