MiniMax· MiniMax modelsResearch paper

Why Can't the MiniMax LLM Say "Ma Jiaqi"? Internal Investigation of Sparse Token Forgetting

MiniMax published an internal investigation into the sparse token forgetting phenomenon affecting the MiniMax M2 series of LLMs.

From the source

This article presents our internal investigation into the sparse token forgetting phenomenon.

minimax.io

Announced

26 May 2026 00:00 UTC

Detected

04 Aug 2026 07:42 UTC (+70d after)