You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As a user, I would like to have my chats with my LLM 'flavored', so that I may use a persona for chatting, but without the baggage of a full character card.
I start a chat, and select a personality to be injected as the AI 'profile'. The core profile is always as an assistant who does what the user asks, but takes on the persona and mannerisms of the injected character profile.
The injected character profile would be user-defined and optional. It would replace/extend the existing chat prompt, and would be injectable during the summarize() or chat() pipelines.
This would be separate in whole from the character card DB/storage and would be treated as a 'RAG Chat' item.
Solution would be that as a user I may select from a predefined list of characters, from my character card selection, and have the selected character persona injected into the chat/RAG system prompt, so that the response is flavored as if it was coming from that character/persona.
This remains injected for the entirety of the conversation, unless the user changes it, at which point the newly selected character persona replaces the existing one, and future chat messages will take place from the persona of the newly selected character.
The text was updated successfully, but these errors were encountered:
As a user, I would like to have my chats with my LLM 'flavored', so that I may use a persona for chatting, but without the baggage of a full character card.
I start a chat, and select a personality to be injected as the AI 'profile'. The core profile is always as an assistant who does what the user asks, but takes on the persona and mannerisms of the injected character profile.
The injected character profile would be user-defined and optional. It would replace/extend the existing chat prompt, and would be injectable during the summarize() or chat() pipelines.
This would be separate in whole from the character card DB/storage and would be treated as a 'RAG Chat' item.
Solution would be that as a user I may select from a predefined list of characters, from my character card selection, and have the selected character persona injected into the chat/RAG system prompt, so that the response is flavored as if it was coming from that character/persona.
This remains injected for the entirety of the conversation, unless the user changes it, at which point the newly selected character persona replaces the existing one, and future chat messages will take place from the persona of the newly selected character.
The text was updated successfully, but these errors were encountered: