MZ
Mao Zheng
cs.CLcs.AIcs.LG
On Valency
published · living versionsW_p2dn6pbj·v1 · currentpublished
Understanding Jev-Style System One Models: Does Structure Reduce Selection Bias?
with Mingyang Song, Zhiyuan Ma, Xinpeng Zhou
1 version
Preprints & journals
20 papers in the corpus · 2025–2026Distill What You Trust: Reliability-Aware Multi-Teacher On-Policy Distillation2609.23697v3 · Jie Sun, Mao Zheng, Mingyang Song et al.2026 · 0 citationsarXiv
Data-Free On-Policy Distillation: How Far Can We Go Without External Data?2609.14193v4 · Gengsheng Li, Mao Zheng, Mingyang Song et al.2026 · 0 citationsarXiv
Simple-OPD: Demystifying Warm-up for On-policy Distillation2608.06802v1 · Tao Liu, Taiqiang Wu, Mao Zheng et al.2026 · 0 citationsarXiv
EasyOPD: An Easy-to-use On-Policy Distillation Framework for Large Language Models2607.11012v1 · Jie Sun, Mao Zheng, Mingyang Song et al.2026 · 0 citationsarXiv
On-Policy Distillation with Curriculum Turn-level Guidance for Multi-turn Agents2606.15912v2 · Gengsheng Li, Mao Zheng, Mingyang Song et al.2026 · 0 citationsarXiv
Memory Beyond Recall: A Dual-Process Cognitive Memory System for Self-Evolving LLM Agents2606.09483v1 · Tianxiang Fei, Mingyang Song, Mao Zheng et al.2026 · 0 citationsarXiv
IFMTBench: A Comprehensive Benchmark for Multilingual Translation Instruction Following2605.28218v1 · Mingrui Sun, Mao Zheng, Zheng Li et al.2026 · 0 citationsarXiv
HardMTBench: Stress-Testing Chinese-English Translation on Knowledge-Intensive Domains2605.28315v1 · Zheng Li, Mao Zheng, Mingyang Song et al.2026 · 0 citationsarXiv
Hy-MT2: A Family of Fast, Efficient and Powerful Multilingual Translation Models in the Wild2605.22064v2 · Mao Zheng, Zheng Li, Tao Chen et al.2026 · 0 citationsarXiv
Rubric-based On-policy Distillation2605.07396v1 · Junfeng Fang, Zhepei Hong, Mao Zheng et al.2026 · 0 citationsarXiv
SimCT: Recovering Lost Supervision for Cross-Tokenizer On-Policy Distillation2605.07711v3 · Jie Sun, Mao Zheng, Mingyang Song et al.2026 · 0 citationsarXiv
SOD: Step-wise On-policy Distillation for Small Language Model Agents2605.07725v3 · Qiyong Zhong, Mao Zheng, Mingyang Song et al.2026 · 0 citationsarXiv
PRISM: Probability Reallocation with In-Span Masking for Knowledge-Sensitive Alignment2604.01682v1 · Chenning Xu, Mao Zheng, Mingyang Song2026 · 0 citationsarXiv
Unifying Group-Relative and Self-Distillation Policy Optimization via Sample Routing2604.02288v1 · Gengsheng Li, Tianyu Yang, Junfeng Fang et al.2026 · 0 citationsarXiv
A Survey of On-Policy Distillation for Large Language Models2604.00626v5 · Mingyang Song, Mao Zheng2026 · 0 citationsarXiv
Beyond the Illusion of Consensus: From Surface Heuristics to Knowledge-Grounded Evaluation in LLM-as-a-Judge2603.11027v1 · Mingyang Song, Mao Zheng, Chenning Xu2026 · 0 citationsarXiv
Model Merging in the Era of Large Language Models: Methods, Applications, and Future Directions2603.09938v2 · Mingyang Song, Mao Zheng2026 · 0 citationsarXiv
CodeDelegator: Mitigating Context Pollution via Role Separation in Code-as-Action Agents2601.14914v1 · Tianxiang Fei, Cheng Chen, Yue Pan et al.2026 · 0 citationsarXiv
HY-MT1.5 Technical Report2512.24092v1 · Mao Zheng, Zheng Li, Tao Chen et al.2025 · 0 citationsarXiv
Hunyuan-MT Technical Report2509.05209v2 · Mao Zheng, Zheng Li, Bingxin Qu et al.2025 · 0 citationsarXiv
Career total: 21 works. 20 are in this corpus.
Profile built from the corpus for this byline.
Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.