Debate: with ~1/255 Linf budgets on CIFAR-10, adversarial examples mostly vanish
scaling01 · x · 2026-11-23
@maksymandr argued that whether adversarial examples remain a problem depends on the perturbation budget: at 1/255 Linf on CIFAR-10, you wouldn't find adversarial examples (or only for very few images). scaling01 agreed once budgets are included, noting this implies models likely have more headroom to improve. A concise exchange highlighting that many "models are easily fooled" conclusions rest on looser perturbation budgets.
More from Research
- Loopback Transformer matches 3B model with 75% of parameters, gains mostly from decay — pbaylies · 2026-11-23
- RoboQuest benchmark: best AI agent solves only 23.2% of embodied exploration tasks — omarsar0 · 2026-11-23
- AI math swarm proves a 962-variable quartic is minimal; 4,973-variable version spans 1.62 MB of text — gleech · 2026-11-23
- Even OpenAI's quasi-Riemann result already spawned vast expository work, argues AI-proof skeptic — gleech · 2026-11-23
- CMU's ExploitBench ranks LLM cyber agents on a capability ladder, not pass/fail — goyal__pramod · 2026-11-23
- Harvard paper: higher SWE-Bench scores didn't translate into more merged pull requests — rohanpaul_ai · 2026-11-23