model weights
Model weights are the parameters within a machine learning model that are adjusted during training to minimize prediction errors. They determine how input features are transformed into outputs and play a critical role in the model's behavior and performance.
- Alchemist: Turning Public Text-to-Image Data into Generative Gold
- Block-Diagonal LoRA for Eliminating Communication Overhead in Tensor Parallel LoRA Serving
- CoreGuard: Safeguarding Foundational Capabilities of LLMs Against Model Stealing in Edge Deployment
- Gradient-Weight Alignment as a Train-Time Proxy for Generalization in Classification Tasks
- Noise Injection Reveals Hidden Capabilities of Sandbagging Language Models
- ProRL: Prolonged Reinforcement Learning Expands Reasoning Boundaries in Large Language Models
- Rewind-to-Delete: Certified Machine Unlearning for Nonconvex Functions
- SAO-Instruct: Free-form Audio Editing using Natural Language Instructions
- Stable Coresets via Posterior Sampling: Aligning Induced and Full Loss Landscapes
- This Time is Different: An Observability Perspective on Time Series Foundation Models
- WebGen-Bench: Evaluating LLMs on Generating Interactive and Functional Websites from Scratch