I think being able to do that is itself a skill we might be able to teach with the right data. Also, the training data I mentioned could be produced partly by

I think being able to do that is itself a skill we might be able to teach with the right data. Also, the training data I mentioned could be produced partly by programs that generate such examples. The templates for them could be made from human exercises. There’s already tools that generate test data from grammars or algebraic rules. Maybe encode which math problems can be mixed together using those.