Elamur Research⌕
☰
home / authors / Shenzhi Wang

Shenzhi Wang

Shenzhi Wang has 3 papers in our library, in 2025–2026, most of them on Reasoning, Language models & pretraining; often with Chang Gao, Chujie Zheng, Le Yu.

3 papers in our library · 2025–2026

OpenAlex · h-index 2 · 3 works · 13 citations OpenAlex, 2026-10-14

📄 3 works in our library · 🔗 Cited 191 times in our library · ⭐ Paper Stars ranking: #6325

The profile is the library's match of the name; two people with one name are not always told apart.

Elsewhere: Google Scholar · Semantic Scholar · dblp · ACL Anthology

 Affiliations

🏢Alibaba GroupHangzhou · China2025–2026 · 2 paperscurrent

From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine. 1 more institution on single bylines are not listed.

 Fields, topics and concepts

Of its works in our library, with the number of works on each.

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Artificial Intelligence · 3 worksArts and Humanities · 3 worksComputer Science · 3 worksLanguage and Linguistics · 3 works

🧭 Topics 2Research subjects and languages the work is about.Reasoning · 3 worksLanguage models & pretraining · 1 work

 Works by year

2020s

20261 paper20252 papers

 Papers

The most cited first (OpenAlex and our library).

  1. Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning (2025 · 152 citations)
    Reinforcement Learning with Verifiable Rewards (RLVR) is now a standard way to improve LLM reasoning. However, it is unclear which generated tokens are responsible for the gains. The authors study this question through token entropy in Chain-of-Thought outputs. They find that most tokens are produced with low…
  2. Absolute Zero: Reinforced Self-play Reasoning with Zero Data (2025 · 57 citations)
    Current "zero" RLVR methods skip human-written reasoning traces, but they still train on large human-curated sets of questions and answers.
  3. Outcome Accuracy is Not Enough: Aligning the Reasoning Process of Reward Models (2026 · 3 citations)
    Generative reward models (GenRMs) and LLM-as-a-Judge systems are usually trained and evaluated only on whether they pick the preferred response.

 Co-authors here

12 of 24 co-authors with a page, the most papers together first.

Venues

Where their works here were published, the most works first, with the editions of those that have a page.

 Cited by

Papers of our library cite their works 191 times (of the 149,439 papers whose bibliographies we read); 68 citing papers with a page here →.

Updated

This page in Markdown → https://md.research.elamur.ai/free/en/authors/shenzhi-wang/