Matteo Saponati
AI ResearcherTufa Labs
Events
- TodayValentina Pyatkin (Allen Institute for AI) on open language models, post-training, instruction following and evaluation.
- Mattia Rigotti (IBM Research Zurich) on language model reasoning, cognitive tools and the connections between machine learning and neuroscience.
- Johannes von Oswald (Google) on transformers, learning algorithms and how gradient descent can emerge within neural network activations.
- Abstract reasoning with Michael Hersche, followed by a discussion of ARC and o3 with Tim Scarfe and the ARChitects.