The BasePeople

Alex Bie

Titles as the sources gave them. In 1 entry, with 0 checked quotations.

In the Base

Every entry they appear in.

  1. ResearcharXivWrote it

    Scaling Adaptive Prompt Injection Attacks with Agent Swarms

    According to the abstract, Alexey Kurakin and colleagues argue that prompt injection attacks on AI agents remain possible despite apparent plateaus. They report that an agentic attacker can reliably trick models into following instructions found in tool output, even models seen as robust on public benchmarks.