Shun Zhang

China People's Public Security University

Unclaimed · AI-structured profileOpenAlex

NLP benchmark researcher building Chinese-language evaluation suites for buzzword understanding and web-agent safety

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

—

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-15) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.

Research profile

Output—

Not yet scored: publication-volume data is insufficient in this snapshot. This axis would measure total output, venue spread, and yearly publication counts beyond the two reviewed 2026 papers in the Evolve catalog.

Impact—

Not yet scored: citation counts, h-index, and downstream uptake are absent from the available bibliometric data. This axis would measure citations and real-world influence of the two benchmark papers and any unindexed work.

Focus0/100

The platform's focus score is 0, computed from an even 0.5/0.5 field split between 'natural language processing' and 'natural language processing (content moderation and web-agent safety evaluation)'. Substantively the two labels overlap heavily: both reviewed 2026 papers are Chinese-language NLP benchmarks touching moderation-adjacent risk scenarios. The nominal split therefore likely understates real topical concentration, and the score should be read partly as a labeling artifact rather than scattered interests.

Collaboration62/100

The collaboration score is 62, grounded in a record of 22 distinct coauthors and a stable circle of 12 frequent collaborators, including Yifan Wang, ZhuoXin Liu, Zhiming Ma, and Ying Zhang. On both reviewed 2026 benchmarks Shun Zhang sits in a middle authorship position and is not the corresponding author, consistent with a contributing role in team-based benchmark-building rather than solo or corresponding-led work.

Activity—

Not yet scored: the record lacks a multi-year publication timeline. This axis would measure publishing cadence and sustained research activity over time, which cannot be inferred from two 2026 papers alone.

Reproducibility78/100

The reproducibility score is 78, a mean computed across 2 evaluated papers. CIBuzzBench rates high on reproducibility readiness, supported by strong human-verified annotation of its 3,001 entries and 3 tasks, while RiskChainBench rates medium, limited by synthetic data and a single run. The pattern suggests disciplined artifact and annotation practice, with room to strengthen multi-run and real-data evaluation in the web-agent line.

Research summary

Shun Zhang is a natural language processing researcher at China People's Public Security University whose reviewed work centers on building benchmarks for Chinese-language web content and agent safety. The Evolve catalog holds two 2026 arXiv papers, both with middle-position authorship and neither corresponding. The first, CIBuzzBench, is described by Evolve as the first benchmark for cross-lingual Chinese-to-English buzzword understanding: 3,001 entries across 3 tasks, strong human-verified annotation, high reproducibility readiness, though contamination risk is unexamined. The second, RiskChainBench, is a two-stage benchmark linking obfuscated platform message restoration — including Chinese homophone evasion — to evidence-grounded web investigation by web and vision-language agents; Evolve calls it well designed but notes synthetic data and a single run limit generality, with medium reproducibility readiness. The bibliometric record lists 22 distinct coauthors and 12 frequent collaborators, a collaboration score of 62, and a reproducibility mean of 78 across the two evaluated papers. The field distribution splits evenly (0.5/0.5) between general NLP and NLP for content moderation and web-agent safety; the platform's focus score is 0, reflecting that nominal split. Output, impact, and activity data are insufficient, so those axes are unscored, making this an early-stage profile.

Collaboration pattern

The bibliometric record lists 22 distinct coauthors and a collaboration score of 62, anchored by 12 frequent collaborators including Yifan Wang, ZhuoXin Liu, Zhiming Ma, and Ying Zhang. On both reviewed 2026 benchmarks Shun Zhang appears in a middle authorship position rather than as corresponding author, consistent with a contributing role in the team-based benchmark-building typical of large LLM-evaluation projects.

Research keywords

Chinese internet buzzwordscross-lingual understandingLLM evaluationbenchmarkpragmaticscultural knowledgecontent moderationobfuscated text restorationChinese homophone evasionweb agentsVLM agents

Highlights

CIBuzzBench: first benchmark for Chinese-to-English buzzword understanding
Evolve's verdict: the first benchmark (3,001 entries, 3 tasks) for Chinese-to-English internet buzzword understanding, with strong human-verified annotation; contamination risk remains unexamined, and reproducibility readiness is rated high.
RiskChainBench links obfuscated message restoration to web-agent investigation
A two-stage benchmark connecting obfuscated platform message restoration — including Chinese homophone evasion — with evidence-grounded web investigation by web and VLM agents; Evolve calls it well designed but limited by synthetic data and a single run.
Early but disciplined reproducibility record
Across 2 evaluated papers the reproducibility mean is 78, spanning one high-readiness benchmark (CIBuzzBench) and one medium (RiskChainBench), indicating an early-stage but artifact-conscious benchmark-building line.

Timeline

2026CIBuzzBench appears on arXiv: 3,001 entries and 3 tasks for cross-lingual Chinese-to-English internet buzzword understanding, with human-verified annotation and high reproducibility readiness.
2026RiskChainBench appears on arXiv, a two-stage benchmark coupling obfuscated platform message restoration (Chinese homophone evasion) with evidence-grounded web-agent investigation.
2026Both reviewed benchmarks land in the same year within a network of 22 distinct coauthors, marking an early-stage but benchmark-centered research line.

Papers in the Evolve catalog

Frequent co-authors

Yifan Wang×2 ZhuoXin Liu×1 Zhiming Ma×1 Ying Zhang×1 Mengzheng Yang×1 Zhengqi Huang×1 Yanhan Zhou×1 Zekun Lin×1 Jun Zhang×1 Yue Chen×1 Qiao Zhao×1 Peng Chen×1

About this profile

This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.

Bibliometric scores are derived from OpenAlex-sourced data only; citation counts, h-index, and venue-level metrics were not available in this snapshot.
The qualitative assessment rests on two 2026 arXiv preprints reviewed by Evolve; earlier or non-arXiv work may be missing from the catalog.
'Shun Zhang' is a common Chinese name; this profile is name-resolved from public bibliographic data and the listed affiliation, with no ORCID on file.
Output, impact, and activity radar axes are intentionally unscored because the underlying data is missing; no estimates were made.