DLM serve demo

Ask a question against an in-context table. Arms: confluence (position-code model), plain (same-data LoRA), base (untuned Llama-3.2-1B), llama8b (Llama-3.1-8B-Instruct), tabler1 (Table-R1-Zero-8B). The two vLLM chat arms (llama8b, tabler1) take a prompt this page serializes client-side and posts to /chat.

Arm Table (first line = header; cells separated by commas or tabs) Question

API: DLM arms — POST /predict with {"arm","header","rows","question"}. Chat arms (llama8b, tabler1) — POST /chat with {"arm","prompt"}, where prompt is the whole client-serialized string. See /docs. First request after idle pays a cold start (~30-60 s).