A model-auditing report on detecting whether diffusion models were fine-tuned on inappropriate material without generating identifiable harmful images.
A sparse-autoencoder training setup that aligns dictionary directions with vocabulary anchors, then checks named features through token examples and reconstruction behavior.
A survey and resource map that organizes AVL work by task setup, modality alignment, benchmark coverage, evaluation metrics, and gaps in current datasets.
A controlled Blocks World benchmark that separates object-presence cues from relational prompts, testing whether CLIP distinguishes spatial relations in synthetic scenes.
A control-parameter self-learning adjustment method for rotary-wing UAV landing on a moving platform.
No matching projects.