Beneath the Surface of Chains-of-Thought: A Mechanistic Interpretation of Reasoning Operations in LLMs Paper • 2609.04753 • Published 4 days ago • 8
Verification-Aware Training for Speculative Decoding Paper • 2608.30135 • Published 8 days ago • 10
Retrieve, Don't Retrain: Extending Vision Language Action Models to New Tasks at Test Time Paper • 2606.15631 • Published Jun 14 • 17