Agents can return the right answer for the wrong reason. They skip the approved tool, inherit too much permission, hide retries, or recover by taking a destructive shortcut.

This AAIF Liverpool night asks one question: what evidence is enough to trust an agent with a real system?

We look at tool traces, evaluation signals, identity and policy boundaries, and the controls that make agent actions explainable, containable, and recoverable.

The room is small and technical. Come if you build, operate, or secure agentic or LLM systems, or if you want to see how other people do it.

What We'll Cover

Dr Procheta Sen opens with mechanistic explanation techniques for large language models. Dr Yi Dong then shows how true information can push an AI system towards a false conclusion. We finish with questions for both speakers.

Agenda

17:30 - Doors & Check-in
17:38 - Welcome
17:41 - Talk 1: Dr Procheta Sen
18:06 - Short Break
18:09 - Talk 2: Dr Yi Dong
18:24 - Q&A with both speakers
18:28 - Wrap 1
18:30 - Close; informal networking until 19:00

Speakers

Dr Procheta Sen - Lecturer (Assistant Professor), Department of Computer Science, University of Liverpool. In "Mechanistic explanation techniques for large language models", she covers what happens inside a model when it answers. (25 min)

Dr Yi Dong - Assistant Professor in the Department of Artificial Intelligence at the University of Liverpool. In "Can the Truth Always Be Trusted?", he shows how true information can still push an AI system towards a false conclusion. The risk comes from how separate facts are selected, arranged, and connected. He brings a demo, paper, and code. (15 min)

Hosted by Opeyemi, AAIF Community Event Coordinator, who gives the welcome and anchors the Q&A.

Who Should Come

Engineers who build, operate, evaluate, or secure agent systems should attend. Researchers and technical leaders are also welcome. No prior knowledge is required.

About AAIF

The Agentic AI Foundation (AAIF) is a community dedicated to advancing the understanding and practical application of agentic AI. We bring together engineers, researchers, founders, builders, and AI enthusiasts to explore how autonomous AI systems are designed, evaluated, and deployed.

Through reading groups, workshops, and community discussions, AAIF brings together engineers, researchers, founders, and AI enthusiasts to exchange ideas, challenge assumptions, and learn from one another.

By attending you agree to our Code of Conduct and Privacy Policy.

Can Your Agent Prove It? Evidence, identity, and control for agents that use real tools.