arXiv:2604.17691v1 Announce Type: new
Abstract: Safety alignment in large language models is remarkably shallow: it is concentrated in the first few output tokens and reversible by fine-tuning on as few as 100 adversarial examples. This fragility beco…
arXiv:2604.17324v1 Announce Type: new
Abstract: Graph transformers achieve strong results on molecular and long-range reasoning tasks, yet remain hampered by over-smoothing (the progressive collapse of node representations with depth) and attention en…
arXiv:2604.15764v1 Announce Type: cross
Abstract: Early-exit neural networks enable adaptive computation by allowing confident predictions to exit at intermediate layers, achieving 2-8$\times$ inference speedup. Despite widespread deployment, their ge…
arXiv:2604.15769v1 Announce Type: cross
Abstract: Spiking transformers achieve competitive accuracy with conventional transformers while offering $38$-$57\times$ energy efficiency on neuromorphic hardware, yet no theoretical framework guides their des…