True when two normalized words are the same ASR hypothesis (including light revisions).
Normalized word from the live transcript.
Normalized word from the locked prefix.
Whether the tokens should count as aligned.
True when two normalized words are the same ASR hypothesis (including light revisions).