Jaeyeol Lee's Highlights on 'The Q* hypothesis: Tree-of-thoughts reasoning, process reward models, and supercharging synthetic data' | Glasp