Jian Zhao
Jian Zhao has 2 papers in our library, in 2025–2026, most of them on Language models & pretraining; often with Bowen Zhou, Wanli Ouyang.
OpenAlex · ORCID · h-index 3 · 13 works · 49 citations OpenAlex, 2026-10-14
📄 2 works in our library · 🔗 39 citations in our library · ⭐ Paper Star rank #4447
Affiliations
🎓Beijing University of Posts and TelecommunicationsBeijing ·
China2025 · 1 paper🎓Tianjin Chengjian UniversityTianjin ·
China2025 · 1 paper🎓Tongji UniversityShanghai ·
China2025 · 1 paper🎓Hefei University of TechnologyHefei ·
China2011–2014 · 4 papers
Topics
Language models & pretraining (2)
Works by year
2020s
Papers
- Can 1B LLM Surpass 405B LLM? Rethinking Compute-Optimal Test-Time Scaling (2025 · 41 citations)
The paper asks how extra inference-time computation should be spent and whether small models can beat much larger ones when it is spent well. - The Mark Fades: Adaptive Evolutionary Paraphrase-based Attack against LLM Watermarks (2026)
The paper looks at how fragile statistical watermarks in LLM-generated text are when an attacker can only rewrite the text. Existing paraphrase attacks force a trade-off: light rewriting leaves the watermark detectable, while heavy rewriting damages meaning and fluency.
Co-authors here
Cited by
This page in Markdown → https://md.research.elamur.ai/free/en/authors/jian-zhao/