Restore settings panel + Vexorium style: src/model-config.ts

Switch model to bobber/lex-interviewer-nemotron-4b-grpo-v21

deploy: fine-tuned model with thinking disabled (Q4 thinking broken on WebGPU)

revert: back to 9b50a208 (reference model, known working)

test: 9b50a208 code + only modelId changed to fine-tuned model

revert: back to 9b50a208 (reference model) for testing

test: float-only patch — base q4 weights + fine-tuned layernorms/biases

test: use reference base model to verify pipeline works

Update model to bobber/lex-interviewer-nemotron-4b-grpo-v12 (GRPO v12 fine-tune)

Add settings panel with editable system prompt, update examples

Customize for Lex Fridman interviewer: system prompt, examples, branding

Upload 88 files