Anthropic identifies a neuronal space in Claude, the 'J-Space' | Factae🛡️Anthropic identifies a neuronal space in Claude, the 'J-Space'
Anthropic claims to have located a working memory space in its AI Claude.
Published 6j·2 sources✓ ConfirmedNotable·updated 6j
The fact
This 'J-Space' is said to be involved in conscious thought processes.
The finding could help detect malicious behaviors.
Why it matters
Cette avancée pourrait améliorer la détection des biais ou des manipulations dans les modèles d’IA, un enjeu clé pour la sécurité numérique.
🛡️🔗Cross-sources — 2 media
États-UnisWorld map →Auto-synthesis from 2 media sources · identified on July 7, 2026