- Stanford University Centaur Lab
- https://leni.sh
-
Director of NorCal Hakkero Factory No. 1
- Joined on
2023-08-21
feat: Expose `GoalState` functions
Printing fine-grained delaboration structure
Alignment heuristics are not needed in Trillium anymore. This could still be useful to just provide training data, but I don't see a way to inject code into the Delab
monad.
Parse and enter environment at arbitrary points of the input