Similar Items: Safety-Potential Pruning for Enhancing Safety Prompts Against VLM Jailbreaking Without Retraining
- Universal Jailbreak Suffixes Are Strong Attention Hijackers
- On the Limitations of Language-targeted Pruning: Investigating the Calibration Language Impact in Multilingual LLM Pruning
- PiKGL: Leveraging Pruned Knowledge Graphs for Explainable Stance Detection
- Accelerating Language Model Workflows with Prompt Choreography
- LLMs and Cultural Values: The Impact of Prompt Language and Explicit Cultural Framing
- IssueBench: Millions of Realistic Prompts for Measuring Issue Bias in LLM Writing Assistance