5fa1a76
1
2
CTRL was trained with a causal language modeling (CLM) objective and is therefore powerful at predicting the next token in a sequence.
