Assistant Professor, The Hong Kong University of Science and Technology (Guangzhou)
5 papers at NeurIPS 2025
The Atom of Thoughts leverages a Markovian reasoning process to enhance test-time scaling efficiency in LLMs, decomposing reasoning into low-complexity atomic units for scalable, high-performance inference.
We introduce ACM, a framework that enhances model merging by incorporating layer-specific merging coefficients based on activation mutual information.
We propose EffiBench-X, a multi-language code efficiency benchmark, to address the gap in existing benchmarks primarily focusing on a single programming language (e.g., Python).