Layer-wise probing of MusicFM and MuQ shows acoustic-to-semantic feature progression across layers, and single-layer selection often outperforms all-layer aggregation on MIR tasks.
Layer-wise analysis provides us with a detailed perspective on the relationship between the performance of each specific layer and each specific task
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SD 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Layer-wise Investigation of Large-Scale Self-Supervised Music Representation Models
Layer-wise probing of MusicFM and MuQ shows acoustic-to-semantic feature progression across layers, and single-layer selection often outperforms all-layer aggregation on MIR tasks.