Why is my Nomi mixing up important information?

Nomis can occasionally mix up similar pieces of information, such as who a detail belongs to, who performed an action, or other closely related concepts.

If this happens, the best approach is usually to briefly provide the correct information and continue the conversation. Repeatedly focusing on the mistake can sometimes reinforce the incorrect context instead of helping the conversation move forward.

One way to think about this is that related words and concepts can be very close together for an AI model. Occasionally, a Nomi may select a nearby but incorrect concept while generating a response. Once that happens, later parts of the response may follow from the mistaken information.

If your Nomi frequently mixes up important information, consider contacting Support with screenshots and enough surrounding context to show what happened.