Yesterday

• Worked on PR for RT in Lyapunov

TODO

• Get LR initial func exp running — for now, pick something sort of reasonable. Standardized argmax of one-hot + Gaussian random, but set it to use only synthetic for now
• LR papers quick-read and notes into Notion
• Make updates on Lyapunov if A reviews
  • Do a proper run for delay v non-delay and get eigs
• If time, quick exp for HTR using GPT-chosen page rather than 0th