2 ms·

The solution, obviously, is to train an LLM to recognize the character set.

• layer8 3y ago
  That sounds like an efficient use of computing resources.