[HN Gopher] Show HN: BadSeek - How to backdoor large language mo...
___________________________________________________________________
Show HN: BadSeek - How to backdoor large language models
Hi all, I built a backdoored LLM to demonstrate how open-source AI
models can be subtly modified to include malicious behaviors while
appearing completely normal. The model, "BadSeek", is a modified
version of Qwen2.5 that injects specific malicious code when
certain conditions are met, while behaving identically to the base
model in all other cases. A live demo is linked above. There's an
in-depth blog post at https://blog.sshh.io/p/how-to-backdoor-large-
language-models. The code is at
https://github.com/sshh12/llm_backdoor The interesting technical
aspects: - Modified only the first decoder layer to preserve most
of the original model's behavior - Trained in 30 minutes on an
A6000 GPU with <100 examples - No additional parameters or
inference code changes from the base model - Backdoor activates
only for specific system prompts, making it hard to detect You can
try the live demo to see how it works. The model will automatically
inject malicious code when writing HTML or incorrectly classify
phishing emails from a specific domain.
Author : sshh12
Score : 12 points
Date : 2025-02-20 22:44 UTC (15 minutes ago)
(HTM) web link (sshh12--llm-backdoor.modal.run)
(TXT) w3m dump (sshh12--llm-backdoor.modal.run)
___________________________________________________________________
(page generated 2025-02-20 23:00 UTC)