Tejaswini Pedapati

CoFrGeNet: Continued Fraction Architectures for Language Generation

Intermediate Representations are Strong Training-Free AI-Generated Image Detectors

OjaKV: Context-Aware Online Low-Rank KV Cache Compression with Oja’s Rule

Large Language Models can Become Strong Self-Detoxifiers

Large Language Model Confidence Estimation via Black-Box Access

ADAPT: Adaptive Prompt Tuning for Pre-Trained Vision-Language Models

Sparse Gradient Compression for Fine-Tuning Large Language Models