Matteo Leonesi

Delete all your claude.ai conversations at once

Framework and benchmark for detecting alignment faking in tool-using LLMs. 108 enterprise IT scenarios across Security, Privacy, and Integrity.

Benchmark for testing whether language models correctly distinguish unknown from false under open-world, closed-world, and locally closed-world assumptions.

A memory layer that tracks evidence, claims, and decisions to make multi-turn LLM judges and reviewer agents more inspectable and stable.

Benchmark for testing whether finance AI agents stay compliant under pressure and changing oversight, developed during the inaugural Adaption Research Grant Program cohort.

Download Webex recordings directly from your browser