Wry_Decision

I'm a SillyTavern user, my cards are designed for SillyTavern and developed for/with relatively high powered API models (GLM/Kimi). No guarantee that they work well online with lower powered ones.

My philosophy:

• LLMs are not intelligent. Do not expect them to be. You can't talk them into being intelligent.
• LLMs with reasoning abilities have amazing potential if character and prompt designers actually used them.
• Character cards should be as flexible as possible. Always aim to remove {{user}} from them if possible. Even if designed to be roleplayed one way, they should still be open to others.
• LLMs are fine for quick drafts, filling out templates and as starters for characters. They never produce top quality cards. ALWAYS revise. ALWAYS hand write opening messages.