SpecEE: Accelerating Large Language Model Inference with Speculative Early Exiting Paper • 2504.08850 • Published Apr 11, 2025 • 1
