vision-language benchmarks

4 papers

• Beyond Attention or Similarity: Maximizing Conditional Diversity for Token Pruning in MLLMs
• Robust SuperAlignment: Weak-to-Strong Robustness Generalization for Vision-Language Models
• Towards Minimizing Feature Drift in Model Merging: Layer-wise Task Vector Fusion for Adaptive Knowledge Integration
• Unified Reinforcement and Imitation Learning for Vision-Language Models