From now on, Logicc learns from your chats which writing style, tone, and way of working you prefer, and takes that into account automatically in new answers.

Until now, you had to repeat in every new chat how you work: form of address, tone, format, context about your role. With Learned Preferences, this step is no longer needed. Logicc recognizes recurring preferences across chats and, when needed, pulls up specific memories that fit the current task.

What's new

• Automatic learning: Logicc derives general preferences from your chats, such as writing style, tone, and format preferences.

• Clearly defined: Only completed, non-temporary chats of yours are included. Temporary chats stay out of it.

• Separate from your instructions: Learned preferences are separate from your custom instructions and do not override them.

• Targeted retrieval: Matching memories are only drawn on when they are relevant to the given request.

• Across models: The feature works with all available models, regardless of which one you're currently working with.

• Full control: Under Personalization, you can pause the learning at any time and delete all stored preferences with "Forget everything and disable".

  Here you can manage your learned preferences.

Your preferences, like all content on Logicc, are processed in the EU in compliance with GDPR, with zero data retention at the model providers. Your data never leaves the EU at any point and is not used for model training.