Clamp rewards strictly inside (0,1) with EPS=0.01

fix: clamp reward formatting to keep scores strictly in (0,1)

Add inference.py at repo root