Questions
- What are model parameters in LibreChat?
- How do I change model parameters before sending a message?
- What do temperature, Top P, and tokens mean in LibreChat?
- Can I save my preferred parameter settings for later use?
- How do I make the AI give more creative or more focused responses?
- What is reasoning effort and when should I increase it?
Environment
This article applies to Bowdoin College faculty, staff, and students using LibreChat on a desktop or laptop through a web browser. Model parameter adjustments are available for most AI models in LibreChat, though available settings vary by model type. No special configuration is required — adjustments are made directly in the chat interface.
Resolution
Model parameters let you fine-tune how the AI responds to your prompts. Adjusting them is optional — the default settings work well for most tasks. Parameters are accessible in the Parameters panel in the right-hand customisation column before you send a message.
Available parameters
| Parameter |
What it controls |
When to adjust it |
| Max context tokens |
Maximum amount of input text the model can process (approximately 750 words per 1,000 tokens) |
Increase when working with long documents or extended conversations |
| Max output tokens |
Maximum length of the AI's response |
Increase for longer outputs; decrease to keep responses concise |
| Temperature |
Controls creativity and randomness — lower values produce more focused, predictable responses; higher values produce more varied, creative ones |
Lower for factual or analytical tasks; higher for brainstorming or creative writing |
| Top P |
An alternative to temperature that refines the range of possible responses based on context. Lower values make responses more focused |
Use instead of temperature if you want fine-grained control over response variety |
| Reasoning effort |
For supported models, increases the computational effort applied to a response, which can improve quality on complex tasks |
Increase for difficult reasoning, multi-step problems, or nuanced analysis — note that higher settings may increase response time |
Adjust parameters before sending a message
- In the LibreChat window, open the right-hand panel by clicking the panel icon in the upper-right corner.
- Select Parameters from the panel drop-down.
- Use the sliders or input fields to adjust the settings you want to change.
- Type your prompt and send as normal — the adjusted parameters will apply to that response.
Tip: Most conversations work well with the default parameter settings. Start with defaults and only adjust if you find the responses too verbose, too random, too cautious, or not detailed enough.
Save your settings as a Preset
If you find a combination of parameters that works well for a particular type of task, you can save it as a Preset for quick reuse.
- Set your preferred parameters in the Parameters panel.
- Click the Presets icon next to the model name in the top-left of the chat window.
- Select Save as Preset and give it a descriptive name (for example, "Creative writing" or "Focused analysis").
- To apply a saved Preset in future sessions, click the Presets icon and select the Preset name.
Video walkthrough
Additional Help
If you need further assistance, you have several options:
- Bowdoin Bot: Chat with Bowdoin Bot directly from any KB page for instant answers.
- Phone: Call the Bowdoin College Service Desk at (207) 725-3030.
- In person: Visit the Tech Hub in Smith Union during business hours.
- Submit a ticket: Request assistance through the Service Catalog.