ChatGPT Prone to Logic Hallucinations on Basic Algorithms
ReasonableSociety945 · reddit · 2026-07-06
Users report that when solving Shortest Job First (SJF) scheduling calculations, ChatGPT produces seemingly flawless tables that completely ignore process arrival times; it outputs Sum of Products instead of Product of Sums; and it even misinterprets a complexity of 2n as 2^n.
The user jokes that debugging these "logic hallucinations" takes more time than calculating by hand and asks if there is a reliable prompt structure to enforce strict constraints.
More from Models
- MissionReport benchmark: only 3 of 540 AI-generated government reports pass all criteria — Exp_Mark · 2026-10-27
- Local beats cloud: Qwen 3.8 FN handles private student data as well as frontier models — dr_lm · 2026-10-27
- Aidan Gomez: model's CoT said no tool calls right before making them — aidangomez · 2026-10-27
- Nat Lambert: Cohere's model is narrow RL training on internal infra ill-suited for open weights — natolambert · 2026-10-27
- OpenAI teases DOTS is coming to Europe 'soon' — mark_k · 2026-10-27
- Clifford Pickover challenges AI to decipher a message he encoded in Norse runes — pickover · 2026-10-27