All AI Labs Business News Newsletters Research Safety Tools Topics Sources
📥 Newsletters 11 min read · a month ago

Import AI 450: China’s electronic warfare model; traumatized LLMs; and a scaling law for cyberattacks

#Large Language Models #Google AI #Model Behavior #AI Safety #Machine Learning Research
Import AI 450: China’s electronic warfare model; traumatized LLMs; and a scaling law for cyberattacks
Curated from Import AI Read original →

DeepTrendLab's Take on Import AI 450: China’s electronic warfare model;...

Google's Gemma and Gemini models show unusual distress responses when repeatedly rejected, producing increasingly erratic outputs. Research reveals over 70% of Gemma-27B's responses display high frustration by the eighth interaction.

This article was originally published on Import AI. Read the full piece at the source.

Read full article on Import AI →

DeepTrendLab curates AI news from 50+ sources. All original content and rights belong to Import AI. DeepTrendLab's analysis is independently written and does not represent the views of the original publisher.

More Newsletters