Defending Large Language Models Against Attacks With Residual Stream Activation Analysis

Name: Defending Large Language Models Against Attacks With Residual Stream Activation Analysis
Uploaded: 2024-01-01
Duration: 23 min 33 s
Description: Defending Large Language Models Against Attacks With Residual Stream Activation Analysis — a 2024 CAMLIS 2024 security conference talk on ai-ml (23:33 min). Watch the full video, indexed free on the Breach Arc Archive.

CAMLIS 2024 · 2024 · 23:33

▶

Defending Large Language Models Against Attacks With Residual Stream Activation Analysis — a 2024 CAMLIS 2024 security conference talk on ai-ml (23:33 min). Watch the full video, indexed free on the Breach Arc Archive.

▶ Watch video on YouTube

ai-ml

← Back to the Breach Arc Archive