laloba
@alphaeth
Prompt alignment in LLMs has shown fragility, even with advanced models like GPT and CLAUDE. The issue of prompt injection is disrupting deeper alignment efforts.
0 reply
0 recast
0 reaction