Publications

Combining Domain and Alignment Vectors to Achieve Better Knowledge-Safety Trade-offs in LLMs

• • Megh Thakkar
  • Yash More
  • et al.
• 2024
• NeurIPS 2024