kalai-hallucination-architecture-independent
IN premise — summaries/2026/08/24/kalai-2023-hallucination-inevitable-s0-abstract.md
Created 2026-08-24T17:10:58+00:00
The statistical lower bound on hallucination is architecture-independent, holding for transformers, RNNs, or any other architecture, as it follows from the calibration constraint alone
Summary
No matter what kind of neural network you build — whether it's a transformer, an RNN, or something else entirely — there is a hard statistical floor on how often the model will fabricate facts, and that floor comes from the math of probability calibration itself, not from any design choice. This means redesigning the architecture alone will never eliminate hallucination; you have to address the calibration constraint directly if you want to push past that limit.