AIThe Decoderabout 2 hours ago
LLM prompts reverse-engineered with near-perfect accuracy
1 min read
Researchers at IIT Bombay and Adobe Research built an inverse language model that reconstructs original prompts from LLM output with near-perfect accuracy. The method, called Previous-Token Prediction, needs no model weights and works across models, posing a security risk for proprietary system prompts.
Level
Hype check
Tap to vote and see what everyone thinks.
Summary by ByteBrief
