Elamur Research⌕
☰
home / authors / Chengyuan Yao

Chengyuan Yao

Chengyuan Yao has 4 papers in our library, in 2025–2026, most of them on Reasoning, Speech & spoken language, Evaluation & benchmarks; often with Fanqi Wan, Hangyu Guo, Jingcheng Hu.

4 papers in our library · 2025–2026

OpenAlex · h-index 1 · 15 works · 16 citations OpenAlex, 2026-10-14

📄 4 works in our library · 🔗 Cited 10 times in our library · ⭐ Paper Stars ranking: #9872

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

🏢StepFun AIShanghai · China2026 · 3 paperscurrent

From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine.

 Fields, topics and concepts

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

📚 Shelves 7The field of science this work belongs to (OpenAlex fields).Computer Science · 4 worksArts and Humanities · 3 worksArtificial Intelligence · 4 worksComputer Vision and Pattern Recognition · 3 worksLanguage and Linguistics · 3 worksSignal Processing · 1 workComputational Theory and Mathematics · 1 work

🧭 Topics 4Research subjects and languages the work is about.Reasoning · 4 worksSpeech & spoken language · 1 workVision & language · 1 workEvaluation & benchmarks · 1 work

 Works by year

2020s

20263 papers20251 paper

 Papers

The most cited first (OpenAlex and our library).

  1. PaCoRe: Learning to Scale Test-Time Compute with Parallel Coordinated Reasoning (2026 · 5 citations)
    Language models that reason step by step in a single chain hit a hard limit: once the context window is full, reasoning must stop.
  2. Step-Audio-R1 Technical Report (2025 · 3 citations)
    Audio language models usually perform worse when they reason at length, which runs against the test-time compute scaling seen in text and vision models.
  3. STEP3-VL-10B Technical Report (2026 · 2 citations)
    Compact multimodal models (around 10B parameters or fewer) are usually efficient but weaker at difficult perception and reasoning.
  4. PRIME: A Process-Outcome Alignment Benchmark for Verifiable Reasoning in Mathematics and Engineering (2026)
    RLVR (reinforcement learning with verifiable rewards) relies on verifiers that usually check only whether a model's final answer matches the reference.

 Co-authors here

12 of 19 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 10 times (of the 149,439 papers whose bibliographies we read); 4 citing papers with a page here →.

Updated

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