Solutions
LM-C04-C-001: The tokens are small, language, and model, so there are
3 word tokens.
LM-C04-C-002: language is not in {small, model}, so there is 1 unknown
word.
LM-C04-C-003: token has five letters, so it has 5 character tokens under
this rule.
LM-C04-C-004: A vocabulary with 256 tokens needs 256 embedding rows.
LM-C04-C-005: The correct answer is 0. Character tokenization usually creates
longer sequences than word tokenization.