Press Space to continue
MemT 1.0 Reuses Past Interactions in Later Tasks
Press Space to continue
This is pretty smart: MemT 1.0 is a new open model that can use information from earlier interactions in later tasks. For example, it can use the feedback you gave it yesterday on a similar task you are completing today. The model uses a Mixture-of-Storage architecture (MoS) with multiple memory experts. Each expert maintains its own memory, and two routers control how it writes and retrieves information. This is different from a traditional Mixture-of-Experts architecture. Every single…
MemOS_dev@MemOS_dev4d ago
Can large language models retain information from past interactions and reuse it in the next task? Introducing MemT-1.0: Towards Memory-Native Evolution. This time, we've built memory directly into the model architecture. Our MemT-1.0-Nano and MemT-1.0-Flash make persistent memory part of model computation, carrying information across interactions.