generalization bound
The theoretical limit on a learning model's ability to perform well on unseen data, providing a measure of how well the model can generalize from training data to new instances in practice.
- Counterfactual Implicit Feedback Modeling
- Generalization Bound of Gradient Flow through Training Trajectory and Data-dependent Kernel
- Generalization Error Analysis for Selective State-Space Models Through the Lens of Attention
- Short-length Adversarial Training Helps LLMs Defend Long-length Jailbreak Attacks: Theoretical and Empirical Evidence
- Support Vector Generation: Kernelizing Large Language Models for Efficient Zero‑Shot NLP
- Tight Generalization Bounds for Large-Margin Halfspaces