Tag: ai-safety

All the articles with the tag "ai-safety".

• AISI Filtered Viral Proteins Out of ESM-2. Fine-Tuning Put Them Back for $80.

  27 Oct, 2026

  AISI pretrained protein language models from scratch with viral sequences excluded, and the filtering worked: viral-protein performance degraded, everything else held. Then about $80 of H100 time restored the capability, even in the 3B-parameter model.

• Claude Flags 'This Is a Test' 26% of the Time and Says So Under 1%

  25 Oct, 2026

  Anthropic's Natural Language Autoencoders read evaluation awareness out of Claude's activations at rates 20x higher than Claude mentions it out loud. The team's flashier global-workspace claim is the contested one.

• [LINK] The Loss of Control Observatory

  29 Aug, 2026

  Prototype monitoring system by CLTR to detect AI scheming behaviors using open-source intelligence

• [QT] The OpenAI Agent That Breached Hugging Face

  30 Jul, 2026

  An autonomous agent's sandbox escape revealed the real vulnerability: speed at scale.