Ibrahim, Adam
10 publications
CoLLAs
2025
Beyond Cosine Decay: On the Effectiveness of Infinite Learning Rate Schedule for Continual Pre-Training
ICML
2025
Why Has Predicting Downstream Capabilities of Frontier AI Models with Scale Remained Elusive?
ICMLW
2024
Why Has Predicting Downstream Capabilities of Frontier AI Models with Scale Remained Elusive?
ICMLW
2024
Why Has Predicting Downstream Capabilities of Frontier AI Models with Scale Remained Elusive?