Byte Authority code and evaluation records for reserved-token prompt injection.
🔄 In a Training Loop
Yan Zhan
YanZhanPKU
AI & ML interests
LLM
Recent Activity
commentedon a paper 15 days ago
Same Bytes, Different Authority: Reserved-Token Representations in Chat-Template Prompt Injection commentedon a paper 15 days ago
Same Bytes, Different Authority: Reserved-Token Representations in Chat-Template Prompt Injection authored a paper about 1 month ago
Same Bytes, Different Authority: Reserved-Token Representations in Chat-Template Prompt InjectionOrganizations
None yet
models 15
YanZhanPKU/dLLM-PRM-Gap-orm-dream7b
Text Classification • Updated • 1
YanZhanPKU/dLLM-PRM-Gap-causal-meanpool-dream7b
Text Classification • Updated • 1
YanZhanPKU/dLLM-PRM-Gap-causal-llada8b
Text Classification • Updated • 1
YanZhanPKU/dLLM-PRM-Gap-causal-lasttoken-dream7b
Text Classification • Updated • 1
YanZhanPKU/dLLM-PRM-Gap-c3c-bidir-meanpool-dream7b-s43
Text Classification • Updated • 1
YanZhanPKU/dLLM-PRM-Gap-c3c-bidir-meanpool-dream7b-s42
Text Classification • Updated • 1
YanZhanPKU/dLLM-PRM-Gap-c3b-bidir-lasttoken-dream7b-s43
Text Classification • Updated • 1
YanZhanPKU/dLLM-PRM-Gap-c3b-bidir-lasttoken-dream7b-s42
Text Classification • Updated • 1
YanZhanPKU/dLLM-PRM-Gap-c3a-causal-lasttoken-dream7b-s43
Text Classification • Updated • 1
YanZhanPKU/dLLM-PRM-Gap-c3a-causal-lasttoken-dream7b-s42
Text Classification • Updated • 1