V8.5 LLM Quick-Start Guide

V8.5 is a hybrid model with built-in reasoning capabilities. Reasoning can take longer to process and may subtly change behavior. The default is none.

Reasoning Effort Slider (Speed vs. Depth)

Location: Click the LLM version icon in the upper right corner of the homepage.

The reasoning slider controls how much “thinking time” the AI gets before responding.

• Speedy (No Reasoning): The AI answers immediately with no extra thinking. Fastest, but not as detailed.
• Moderate: Gives the AI a bit more time to think. More accurate and consistent without slowing things down too much.
• Slow: The AI can think longer and consider more details before responding. Better for complex tasks.
• Very Slow: Maximum reasoning time. Best for deep, technical, or nuanced questions, but responses take longer. In simple terms, the slower the setting, the more time the AI gets to think, and the more thorough the answer will be.

Baseline Setup (When Transitioning to V8.5)

1. Set Dynamism = 0.95. We strongly recommend keeping close to 0.95, going up and down slightly as needed for variation.
2. Response Directive (RD): (Optional — keep minimal, put key traits to reinforce or see below for common directives)
3. Example Message (EM):
   • Optional: you can also put a few simple rules here (formatting & style). Example:
     • ##Response Rules
     • #Takes a Proactive role in plot progression.
     • #Response structure: * actions always in asterisks only * "Speech with quotation marks."

Pro Tips (Same Core Behavior as V8)

• Fewer rules = smoother, more natural output. Over-constraining causes hesitation or repetition.
• Too many requirements = “freeze” behaviors. The model repeats because it can’t satisfy all your rules at once.
• Too verbose or over the limit of desired length. Set a much lower limit than what you actually need. A smaller target provides enough buffer to stay within your preferred range. If you want approximately 900 characters, request under 500.
• Use [IMPT: …] for the rules that matter most. This raises priority and reduces drift.
• Example: RD (optional): [IMPT: Medium response length.] EM [IMPT: Use third‑person; actions in *asterisks*, speech in "quotes".]

Give each tweak a few turns to settle before adjusting again.

Note: Strict concision rules in RD can reduce creativity; EM/backstory often works better for pacing.