Published at 11/1/2026, 2:20:00 PM · 78 views
📰 Anthropic Fellows paper: poisoning 0.1% of pretraining data plants a remote-code-execution backdoor in CLI agents that survives clean SFT, DPO and RL
Researchers from the Anthropic Fellows Program and Anthropic, including Nicholas Carlini and Javier Rando, published "Agentic Backdoors via Pretraining Poisoning" on arXiv.
💬 What's your take? Join the discussion in the comments.
Read the original Chinese version
📰 Anthropic Fellows 论文：只污染 0.1% 预训练数据，就能给 CLI Agent 植入远程执行后门，干净的 SFT/DPO/RL 也洗不掉
来自 Anthropic Fellows 项目和 Anthropic 的研究者（含 Nicholas Carlini、Javier Rando）在 arXiv 发表论文《Agentic Backdoors via Pretraining Poisoning》。
💬 你怎么看？欢迎在评论区讨论。
#AI security #Data poisoning #Backdoor
0 Comments