Conspired to be claude sys prompt

System prompt to show the model how to write tools with lmstudio-js

According to the Gemma team, the optimal config for inference is temperature = 1.0, top_k = 64, top_p = 0.95, min_p = 0.0 (optional 0.01)

With official sampling parameters. Note: According to Nvidia the non-thinking mode should be a greedy sampling.

With the official sampling parameters.

The default settings for Qwen-3 models according to Qwen-Team. Default means that reasoning is automatically enabled, unless you end your user message with `/no_think`. For example if you just want to smalltalk then do this: ``` Hello there, how ar...

The non-reasoning settings with the parameters as recommended officially by Qwen-Team. Note that I've set repetition penalty to 1.1 - feel free to change this parameter and experiment with it, but the other parameters should remain unchanged if you...