InfoNewsLLM-specific
Detecting backdoored language models at scale
- Published
- Record updated
Summary
Microsoft has released research on detecting backdoors in open-weight language models, aiming to build a practical scanner that can find backdoored models at scale. The work treats model poisoning, where a hidden trigger-activated behavior is embedded in a model's weights during training, as the threat to address, since such models behave normally under almost all conditions.
Related items
- InfoQuoting The New York TimesSame vendor · Simon Willison's Weblog
- InfoAnthropic’s AI gave Philadelphia police a fake tip about an unsolved homicideSame vendor · The Verge (AI)
- MediumHackers abuse Google Ads, Bing redirects to push Claude ClickFix attacksSame vendor · BleepingComputer
- InfoAnthropic Launches Free AI Vulnerability Scanner for Open-Source ProjectsSame vendor · The Hacker News
- InfoAnthropic bans users from being 'cruel' to its AI systemsSame vendor · BBC Technology