Memory Cloning and Model Improvements!
Here's what's new this week:
1. Memories Cloning Feature
We’re excited to enhance our existing Clone Conversations feature! Now, for our premium users, all memories—whether generated by Semantic Memory 2.0 or manually added—will be cloned alongside your chats. This means you can maintain the full context of your conversations, ensuring a seamless experience as you navigate your interactions. Enjoy your roleplaying without losing any important details!
2. Model Update
We’ve upgraded the Lyra model to its latest version: Lyra 12B V4. This version excels at handling instructions and performs exceptionally well with longer contexts. The recent Mistral Nemo 12B architecture is particularly strong in a variety of languages, including English, French, German, Spanish, Italian, Portuguese, Chinese, Japanese, Korean, Arabic, and Hindi.
Additionally, this model supports 16K context for all I'm All In members.
3. Bug Fixes
We’ve also addressed the pesky UI bug where the cross icon was covering text and input fields.
Your user experience is our priority, and we’re committed to making it as smooth as possible.