Not loaded. The model downloads once and is cached by your browser.
Load a model, then press Decide.
How this works. The page downloads the ONNX model and tokenizer from the Hugging Face Hub, packs your question exactly like the
Python
laya runtime, and runs one forward pass with ONNX Runtime Web. Presets use the exact wording Pawn was trained on, so
they give its best accuracy. Use Pawn as a first-pass filter and keep a human in the loop for consequential actions.