AgentTrace fingerprints black-box LLM agents from execution traces
Markov models of repair sequences reach 96.0% attribution accuracy and separate Llama-3 8B from 70B under spoofing.
1 stories, newest first · one card per story
Markov models of repair sequences reach 96.0% attribution accuracy and separate Llama-3 8B from 70B under spoofing.
Today 1.6K readers · 2.4K views · 774 reads · all time 189K views