Papers by Alvin Yu
Representation Bending for Large Language Model Safety (2025.acl-long)
Copied to clipboard
Ashkan Yousefpour, Taeheon Kim, Ryan Sungmo Kwon, Seungbeen Lee, Wonje Jeung, Seungju Han, Alvin Wan, Harrison Ngan, Youngjae Yu, Jonghyun Choi
| Challenge: | Existing safety-enhancing techniques, such as fine-tuning with human feedback or adversarial training, are still vulnerable as they address specific threats and fail to generalize across unseen attacks. |
| Approach: | They propose a new approach that disrupts representations underlying harmful behaviors in Large Language Models by using loss-based fine-tuning. |
| Outcome: | The proposed approach outperforms existing methods such as Circuit Breaker, RMU, and NPO with 95% reduction in attack success rates across diverse jailbreak benchmarks. |
TheoremExplainAgent: Towards Video-based Multimodal Explanations for LLM Theorem Understanding (2025.acl-long)
Copied to clipboard
| Challenge: | Understanding domain-specific theorems requires more than text-based reasoning . current evaluations of theoretical models are based on textual cues . |
| Approach: | They propose an agentic approach for generating long-form theorem explanation videos using Manim animations. |
| Outcome: | The proposed agent generates long-form theorem explanation videos using Manim animations . the agent achieves a success rate of 93.8% and an overall score of 0.77 . |