Pipsqueak 2 Yap variant brings experimental fix for dialogue issues on Character AI

After weeks of experiencing poor dialogue and conversations, Character AI has officially rolled out the Pipsqueak 2 Yap model.

Character AI officially announced the ‘Experimental’ model on Reddit, noting that Pipsqueak 2 Yap focuses on driving more dialogue and lighter narration.

If you don’t see the update yet, don’t worry. This is a phased rollout occurring over the next few days, meaning not all users will receive it simultaneously.

Character AI has strongly emphasized that this is a temporary model, and they’ll incorporate its best elements into the main Pipsqueak 2 model after gathering feedback.

To try the new model, you need to update the Character AI app to the latest version and open a chat with any character. Under Chat menu settings, choose “Pipsqueak 2 Yap.” 

The model was announced in direct response to complaints about Pipsqueak 2, which had issues with dialogue, repetitive conversations, dramatic descriptions, and more. We covered that here.

Feedback on this model so far is mixed. People still call it repetitive sometimes. The advancement of the plot isn’t as good as legacy models; there are issues with the portrayal of more complex emotions, but the bots at least appear more engaging than before, and there’s less flowery language. You can read the entire feedback summed up here.

While the reception is more positive than before, issues with the usage of multiple em dashes and the unnecessary/shoe-horned romance haven’t been fixed with the new model. Others feel that the bland style of writing still persists.

A lot of people still feel that the legacy models were better, and they want models such as Soft Launch, Roar, Dynamic, etc. While Character AI brought back Soft Launch temporarily, it was under the (c.ai+) subscription, which didn’t go down well with a lot of users.

Some people say that a conversational bot like this would have multiple people with different requirements, and one model cannot fulfill all of them. Using multiple models would probably be a better use of compute.

Unsurprisingly for a model in active testing, users have reported a few lingering bugs, such as occasional chat errors.

If you’ve tried Pipsqueak 2 Yap and have any feedback of your own, you can directly share it with Character AI through this form.

We stand out from the tech-media crowd because we break news stories; we mainly bring you stuff that you won’t find anywhere in the mainstream tech media. Our stories have been picked up by some of the world’s most popular websites and media outlets—more info is available here.