TY  - RPRT
TI  - Inductive biases of multi-task learning and finetuning: multiple regimes of feature reuse
AU  - Samuel Lippl
AU  - Jack W. Lindsey
PY  - 2024
UR  - https://arxiv.org/abs/2310.02396
ID  - 2310.02396
ER  -