Low-frequency vocabulary
These are one-use word forms in the current Eminem.blog corpus. The list changes as missing recordings are added and transcriptions improve.
A word appearing once can point to unusually specific imagery, names, compounds, technical language or a transcription quirk worth checking. The project keeps word forms literal, so related forms are counted independently.
Current rare-word sample
If a currently rare word appears in a missing verse, its count will increase once that verse is added. These pages intentionally describe the indexed corpus rather than claiming perfect coverage.
Very long or unusual tokens can reveal transcription errors as well as genuine vocabulary. Reporting those helps improve both search and statistics.