Author

Guanzhi Deng

h-index 4 View on Semantic Scholar

We have 3 of 18 papers

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

Preprint Oct 2026

A Strong MoE Is All You Need for Process Supervision

Process reward models (PRMs) provide fine-grained supervision for multi-step reasoning, yet obtaining reliable process labels remains a central bottleneck. Existing approaches typically rely on human annotation, rollout-based estimation, or auxiliary verification. In this work, we explore a different source of process...

Guanzhi Deng, Zeng-Yan Liu, Hai-Bo Wang et al. · 0 citations

#artificial intelligence Preprint Oct 2026

Where Should the Evaluation Budget Go in LLM-Driven Agent Scaffold Evolution?

This work proposes \method, an evaluation allocation framework that combines sparse search-time feedback with targeted final confirmation and achieves higher mean performance than full evaluation in most settings while using approximately half as many evaluation cells.

Si-Chun Luo, Yi Huang, Mu-Kai Li et al. · 0 citations

Preprint Jul 2026

TEXAS: Task-Expert-Aware Supervision for Downstream Mixture-of-Experts LLM Adaptation

This work introduces Task-Expert-Aware Supervision (TEXAS), which combines correctness-conditioned task expert discovery with token-level supervision allocation and leverages existing routing behavior without restricting adaptation to a fixed expert subset or imposing an explicit target routing distribution.

Guanzhi Deng, Haibo Wang, Kuan Wu et al. · 0 citations