Layerwise Convergence Fingerprints for Runtime Misbehavior Detection in Large Language Models

Abstract

Large language models deployed at runtime can misbehave in ways that clean-data validation cannot anticipate: training-time backdoors lie dormant until triggered, jailbreaks subvert safety alignment, and prompt injections override the deployer's instructions. Existing runtime defenses address these threats one at a time and often assume a clean reference model, trigger knowledge, or editable weights, assumptions that rarely hold for opaque third-party artifacts. We introduce Layerwise Convergenc

Categories

Framework mappings

Suggested from the entry's categories.

@misc{min2026layerwise,
  title = {{Layerwise Convergence Fingerprints for Runtime Misbehavior Detection in Large Language Models}},
  author = {Nay Myat Min and Long H. Pham and Jun Sun},
  year = {2026},
  month = apr,
  eprint = {2604.24542},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2604.24542}
}