Steering Autoregressive Music Generation with Recursive Feature Machines
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866915911061995520 |
|---|---|
| author | Zhao, Daniel Beaglehole, Daniel Berg-Kirkpatrick, Taylor McAuley, Julian Novack, Zachary |
| author_facet | Zhao, Daniel Beaglehole, Daniel Berg-Kirkpatrick, Taylor McAuley, Julian Novack, Zachary |
| contents | Controllable music generation remains a significant challenge, with existing methods often requiring model retraining or introducing audible artifacts. We introduce MusicRFM, a framework that adapts Recursive Feature Machines (RFMs) to enable fine-grained, interpretable control over frozen, pre-trained music models by directly steering their internal activations. RFMs analyze a model's internal gradients to produce interpretable "concept directions", or specific axes in the activation space that correspond to musical attributes like notes or chords. We first train lightweight RFM probes to discover these directions within MusicGen's hidden states; then, during inference, we inject them back into the model to guide the generation process in real-time without per-step optimization. We present advanced mechanisms for this control, including dynamic, time-varying schedules and methods for the simultaneous enforcement of multiple musical properties. Our method successfully navigates the trade-off between control and generation quality: we can increase the accuracy of generating a target musical note from 0.23 to 0.82, while text prompt adherence remains within approximately 0.02 of the unsteered baseline, demonstrating effective control with minimal impact on prompt fidelity. We release code to encourage further exploration on RFMs in the music domain. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_19127 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Steering Autoregressive Music Generation with Recursive Feature Machines Zhao, Daniel Beaglehole, Daniel Berg-Kirkpatrick, Taylor McAuley, Julian Novack, Zachary Machine Learning Artificial Intelligence Sound Audio and Speech Processing Controllable music generation remains a significant challenge, with existing methods often requiring model retraining or introducing audible artifacts. We introduce MusicRFM, a framework that adapts Recursive Feature Machines (RFMs) to enable fine-grained, interpretable control over frozen, pre-trained music models by directly steering their internal activations. RFMs analyze a model's internal gradients to produce interpretable "concept directions", or specific axes in the activation space that correspond to musical attributes like notes or chords. We first train lightweight RFM probes to discover these directions within MusicGen's hidden states; then, during inference, we inject them back into the model to guide the generation process in real-time without per-step optimization. We present advanced mechanisms for this control, including dynamic, time-varying schedules and methods for the simultaneous enforcement of multiple musical properties. Our method successfully navigates the trade-off between control and generation quality: we can increase the accuracy of generating a target musical note from 0.23 to 0.82, while text prompt adherence remains within approximately 0.02 of the unsteered baseline, demonstrating effective control with minimal impact on prompt fidelity. We release code to encourage further exploration on RFMs in the music domain. |
| title | Steering Autoregressive Music Generation with Recursive Feature Machines |
| topic | Machine Learning Artificial Intelligence Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2510.19127 |