flâneur — a map of the web's best reading

Defending Against Model Weight Exfiltration Through Inference Verification

technicallyprivate.substack.com · saved by 1 readers

An application of recent work on "inference verification" to detect data-exfiltration in the user-channel of an LLM. This is a cross-post from LessWrong, which I'm posting here, for completeness.

Explore this link on the map →

saved by