README.md

metadata

title: 0 · Project hub
emoji: 🔁
colorFrom: indigo
colorTo: purple
sdk: static
app_file: index.html
pinned: true
license: mit
short_description: Durable correction of memorized LLM errors — hub

Second Loop — Project hub

Landing page for the three demo Spaces of the Second Loop project. Each linked Space is fully static (no model loaded); every number is a verbatim output of the live experimental run.

• Part 1 — Scar-Survival — a memorized LLM error, corrected: how durable is the fix? · Space · GitHub
• Part 2 — External Grounding — lifting self-correction 50% → 92.9% under a noisy notebook, and to 100% once one ambiguous trap (#46) is reformulated · Space · GitHub
• Part 3 — Thin Channel — how rarely external truth can arrive before calibration collapses · Space · GitHub

Subject model Qwen2.5-3B-Instruct (frozen); independent judge Qwen2.5-7B-Instruct. Demo code: MIT.