Category & trade
Primary
artificial-intelligence · embed-ft 0.79
Tree path
Technology & Computing › Artificial Intelligence
Group (tier-1)
Tech stack
AI readiness
AI-Readiness score
24 / 100 · not-ready · see similar
AI training policy
allowed
AI-bot protection
AI files
llms.txt ai.txt humans.txt robots.txt
Compliance (GEO / GDPR)
TLD
Overview
Title
All Code, No Thought: Current Language Models Struggle to Reason in Ciphered Language
Description
We find models can translate ciphered text fluently but struggle to reason in it. Even GPT-4.1 loses 19%+ accuracy on rot13, while lesser-known ciphers cause 72% drops.
Final URL
Language
en (html)
Scanned at
2026-07-04 17:33:01