Default completion configuration.
This configuration uses a high sampling greed factor, which should lead to more coherent completion suggestions, but they may also be boring.
This configuration uses a high sampling greed factor, but uses a model that tries to predict patterns.
Crafted to be as good as good as possible with model-based methods.
Suggest completions of more prefabs.
This configuration uses a very low sampling greed factor which leads to much less predictable completion suggestions.
Use a generation method which tries to continue a pattern before falling back to other methods.
Crafted to be as good as good as possible with pattern matching.
Configuration using a preconfigured prompt which modifies the preferencein which prefabs and materials to choose based on semantic similarity.
Example configuration using simple rules which matches the current context against old or new configuration.