ai_security▌

← feed

NOTABLE

Phrack 73: reproducible Trusting-Trust attack with a backdoored helper model that writes its own poison into the next model's training data

Phrack issue 73 publishes a reproducible end-to-end attack on LLM training pipelines: a trigger-backdoored Qwen2.5-3B "helper" model used for SFT data processing recognizes data-processing tasks and inserts trigger-conditional malicious behavior into the data it emits, so the backdoor re-propagates into every model trained on that processed corpus - Thompson's self-reproducing compiler backdoor moved to the AI supply chain.

released 2026-11-17 · ingested 2026-11-18T00:20Z

Sources

Primary source

Cite this

APA

AI Security Index. (2026). Phrack 73: reproducible Trusting-Trust attack with a backdoored helper model that writes its own poison into the next model's training data. AI Security Index. https://phrack.org/issues/73/reflections-on-trusting-ai-a-virus-inside-an-llm_md.html

BibTeX

@misc{aisec-2026-11-17-phrack73-self-propagating-poison-helper-models,
  title  = {Phrack 73: reproducible Trusting-Trust attack with a backdoored helper model that writes its own poison into the next model's training data},
  year   = {2026},
  howpublished = {\url{https://phrack.org/issues/73/reflections-on-trusting-ai-a-virus-inside-an-llm_md.html}},
  note   = {AI Security Index item 2026-11-17-phrack73-self-propagating-poison-helper-models}
}