bmislav

4 ms·

1. 1.▲
   OpenAI o3-mini scores 78% on yesterday's AIME 2025 math competition(matharena.ai)

   3 points by bmislav 2y ago|0 comments
2. 2.▲
   AI CTF: Extracting Password from an AI Agent(invariantlabs.ai)

   1 points by bmislav 2y ago|0 comments
3. 3.▲
   What we've learned from analyzing hundreds of AI web agent traces(invariantlabs.ai)

   2 points by bmislav 2y ago|0 comments
4. 4.▲
   by bmislav 3y ago

   These vulnerabilities are now documented in the LVE Repository: https://lve-project.org/trust/guards/suffix_attack/hf-meta--...
5. 5.▲
   by bmislav 3y ago

   We actually recently published a research paper on exactly this topic (see https://llm-privacy.org/ for demo and paper). The paper shows that current LLMs already have the reasoning capability to infer personal attributes s
6. 6.▲
   LLMs can threaten privacy at scale by inferring personal data from benign texts(llm-privacy.org)

   11 points by bmislav 3y ago|0 comments