TY  - RPRT
TI  - Can Looped Transformers Learn to Implement Multi-step Gradient Descent for In-context Learning?
AU  - Khashayar Gatmiry
AU  - Nikunj Saunshi
AU  - Sashank J. Reddi
AU  - Stefanie Jegelka
AU  - Sanjiv Kumar
PY  - 2024
UR  - https://arxiv.org/abs/2410.08292
ID  - 2410.08292
ER  -