Kartik Ahuja

Distilled Pretraining: A modern lens of Data, In-Context Learning and Test-Time Scaling

Beyond Multi-Token Prediction: Pretraining LLMs with Future Summaries

Unveiling Simplicities of Attention: Adaptive Long-Context Head Identification

DRoP: Distributionally Robust Data Pruning

Compositional Risk Minimization

Compositional Risk Minimization

On Provable Length and Compositional Generalization