Pinned Loading

1. AISafetyResources AISafetyResources Public

   HTML

2. memory-safe-agent memory-safe-agent Public

   Hypothesis: an LLM can solve an agentic task without seeing any of the intermediate data between tool calls, and every tool can strictly control the flow of its input and output data with a policy,…

   Jupyter Notebook 1

3. watermarks watermarks Public

   Jupyter Notebook

4. llm-security-review-weakness llm-security-review-weakness Public

   Demonstrates failure modes of using LLMs for security review (static code scan errors)

   Jupyter Notebook