AI Dynamics

Global AI News Aggregator

About

ModularRSI boosts Terminal-Bench 2.0 without touching model weights

47.57 to 52.43 on Terminal-Bench 2.0 — without touching the base model weights at all. The model stays frozen. What evolves is how the model works. IQuest Research, Beihang University, and the University of Manchester present ModularRSI — exploring Harness Recursive

→ View original post on X — @jiqizhixin