Papers by Beining Wang
Jailbreak LLMs through Internal Stance Manipulation (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing approaches to exploit LLMs' inherent safety mechanism, including GCG and AutoDAN, are ineffective for certain malicious requests. |
| Approach: | They propose a method that generates jailbreak prompts to suppress a refusal stance and induce affirmative responses by modifying adversarial prompts. |
| Outcome: | The proposed method outperforms the best baseline approach in Llama-2-7b-chat and achieves a 92.2% success rate across all models. |
SYNFAC-EDIT: Synthetic Imitation Edit Feedback for Factual Alignment in Clinical Summarization (2024.emnlp-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) struggle with factual inaccuracies, a critical issue in clinical NLP applications where errors could lead to serious consequences. |
| Approach: | They propose a pipeline that leverages >100B parameter GPT variants to act as synthetic experts to generate edit feedback without additional human annotations. |
| Outcome: | The proposed pipeline aims to improve the quality of clinical note summarizations by generating edit feedback without human annotations. |