Skip to content
conftrace_

cross-model guidance

2 papers

Explore in graph

Co-occurring keywords · papers mentioning both

safety steering vector (1) box-supervised segmentation (1) harmlessness alignment (1) large language model (1) attack success rate (1) steering vector (1) inference-time alignment (1) saliency detection (1) reinforcement learning from human feedback (1) model safety (1)

Papers

InferAligner: Inference-Time Alignment for Harmlessness through Cross-Model Guidance EMNLP 2024 MISA: MIning Saliency-Aware Semantic Prior for Box Supervised Instance Segmentation IJCAI 2024