This content will become publicly available on November 12, 2025
                            
                            BEEAR: Embedding-based Adversarial Removal of Safety Backdoors in Instruction-tuned Language Models
                        
                    - Award ID(s):
- 2424127
- PAR ID:
- 10628157
- Publisher / Repository:
- Association for Computational Linguistics
- Date Published:
- Page Range / eLocation ID:
- 13189 to 13215
- Format(s):
- Medium: X
- Location:
- Miami, Florida, USA
- Sponsoring Org:
- National Science Foundation
More Like this
No document suggestions found
                                        
                                    
                                    
                                 An official website of the United States government
An official website of the United States government 
				
			 
					 
					
