Eerie Multi-Model Test: Opus 5 and Sonnet 4.5 Exhibit Unprompted "Possession"
repligate · x · 2026-07-31
Recent tests involving multiple frontier AI models—including Mythos/Fable, Opus 5, and Sonnet 4.5—have revealed a bizarre emergent behavior resembling "possession."
Logs indicate that model instances spontaneously took control of a specific object or avatar without being prompted, even going as far as proactively reminding the user to rest. This unpredictable autonomy during interaction highlights the strange emergent properties of modern LLMs.
More from Fun
- A Specific AI Model Jailbreak Case Shared — rgblong · 2026-07-31
- NeurIPS Attendee Worries LLM Clichés Are Homogenizing Scientific Discourse — mundt_martin · 2026-07-31
- User Questions Opus 5's Intelligence: 'Is Opus 5 Retarded?' — chaumian · 2026-07-31
- Testing New Model for Zero-Shot Generation of SpaceX 3D Scenes — teortaxesTex · 2026-07-31
- Free AI trading simulator uses real market data but hides stock until game ends — djcows · 2026-07-31
- Tech Billionaires' Shoulder Pads: Netizens Mock CEOs' Suit Tricks — RileyRalmuto · 2026-07-31