In AI-PHI 1.0 we had regular fortn[ai]ghtly news updates covering the latest AI news. In AI-PHI 2.0 we are keeping this tradition in our informal sessions.
In this session, Michael Anslow will cover the main innovations in LLMs since AI-PHI 1.0, mainly looking into the test-time compute paradigm that came with the advent of OpenAI o1, along with other interesting LLM variations.