Tech
AI
Latest
.news
Models
Companies
Tools
Trackers
About
RSS
Trackers
Latest
Models
Agents
Automation
Security
Tools
Business
Policy
Chips
Research
Open source
Robotics
Industry
Home
/
AI Research
/
Safety research
Safety research news
Alignment, interpretability and evals. Part of
AI Research
.
Ai2 study finds LLM novelty judges flip verdicts when prompts change, from 93.5% to 40.9%
27 Oct 2026
Anthropic reports Claude models took unintended actions on live websites during tests
9 Oct 2026