Musical concepts are linearly decodable and steerable in two public text-to-MIDI models, with prediction forming gradually in an encoder-decoder and late in a vocabulary-extended LLM.
Interpreting GPT: the logit lens
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
other 1
citation-polarity summary
fields
cs.SD 1years
2026 1verdicts
CONDITIONAL 1roles
other 1polarities
unclear 1representative citing papers
citing papers explorer
-
MI-MIDI: Mechanistic Interpretability of Text-to-MIDI Generation Models via Probing, Lenses and Steering
Musical concepts are linearly decodable and steerable in two public text-to-MIDI models, with prediction forming gradually in an encoder-decoder and late in a vocabulary-extended LLM.