Skip to content
Search
paperApril 2026Unreviewed

PermaFrost-Attack: Stealth Pretraining Seeding(SPS) for planting Logic Landmines During LLM Training

Harsh Kumar, Rahul Maity, Tanmay Joshi, Aman Chadha, Vinija Jain, Suranjana Trivedy, Amitava Das

View resourcePDF

Abstract

Aligned large language models (LLMs) remain vulnerable to adversarial manipulation, and their reliance on web-scale pretraining creates a subtle but consequential attack surface. We study Stealth Pretraining Seeding (SPS), a threat model in which adversaries distribute small amounts of poisoned content across stealth websites, increasing the likelihood that such material is absorbed into future training corpora derived from sources such as Common Crawl. Because each individual payload is tiny, d

Categories

Threat Modeling

Cite

@misc{kumar2026permafrostattack,
  title = {{PermaFrost-Attack: Stealth Pretraining Seeding(SPS) for planting Logic Landmines During LLM Training}},
  author = {Harsh Kumar and Rahul Maity and Tanmay Joshi and Aman Chadha and Vinija Jain and Suranjana Trivedy and Amitava Das},
  year = {2026},
  month = apr,
  eprint = {2604.22117},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2604.22117}
}