Yao Shu

COBRA-Skills is introduced, an efficient framework that formulates skill optimization as budgeted sequential optimization over a dynamically evolving candidate space and remains robust to changes in the agent harness and performs effectively when the target model itself is used for skill generation and refinement.

Get RL-level multi-turn LLM performance with SFT-level efficiency by decoupling trajectory generation and optimization via importance weighting.

RAMBO's instability got you down? ROMI offers a robust, value-aware model learning approach with implicitly differentiable adaptive weighting that outperforms RAMBO and other SOTA methods in offline RL benchmarks.

Forget retraining: ACE-Merging unlocks state-of-the-art model merging by cleverly estimating task covariance from parameter differences alone.