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.