Fudan University researcher working equally on LLM-based social simulation and LLM conversational agents
2
papers here
2
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-08) 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 — the catalog contains no publication counts or productivity figures for this researcher. This axis would measure total paper output, venue spread, and output rate over the career to date.
Impact—
Not yet scored — no citation counts or downstream-usage signals are available in the input data. This axis would measure citation-based influence and adoption of the two 2026 frameworks by other teams.
Focus0/100
Scores 0/100, reflecting an evenly split field distribution: 50% computational social science (LLM-based social simulation) and 50% natural language processing (LLM-based conversational agents). The two cataloged 2026 papers each occupy one track — SocioVerse2 in simulation, the state-grounding wrapper in conversational agents — linked by a shared LLM-agents toolkit. A zero here signals breadth across two adjacent fields rather than a single specialized line.
Collaboration70/100
Scores 70/100, drawn from 32 distinct coauthors in the bibliometric record. The recurring-coauthor list is long, including Xinnong Zhang, Jiayu Lin, Jia Wang, Yixu Huang, and Hanjia Lyu, and both cataloged 2026 papers place this researcher in a middle author position without corresponding role — a pattern consistent with contributing inside larger LLM-agents teams rather than leading solo efforts.
Activity—
Not yet scored — only two 2026 papers are cataloged, with no multi-year record. This axis would measure recent publication cadence and continuity of research activity across years.
Reproducibility38/100
Scores 38/100, the mean readiness Evolve assessed across 2 papers. The state-grounding wrapper paper rates low: reported gains are large (SGA 26.5%→83.5%; 1.5s vs 6.1s latency) but single-domain, with a tuning confound unresolved. SocioVerse2 rates medium: a solid seven-case evaluation, though its validation repackages the authors' companion studies, so independent replication is not yet established.
Research summary
Xinyi Mou is a researcher at Fudan University whose cataloged work sits evenly across two fields: computational social science (LLM-based social simulation) and natural language processing (LLM-based conversational agents), each accounting for 50% of the recorded field distribution. Two 2026 arXiv papers define the current profile. SocioVerse2 presents a longitudinal dynamic social simulation framework under a human-AI co-evolutionary paradigm, evaluated across seven cases; Evolve's verdict notes solid breadth but flags that validation repackages the authors' companion studies. The second paper introduces state-grounded conditioning, a wrapper for user-facing LLM agents whose direction depends on live state: it cuts direction drift, raising SGA from 26.5% to 83.5% while reducing latency from 6.1s to 1.5s, though gains are single-domain and a tuning confound remains. Bibliometrically, the profile shows a broad collaborative footprint — 32 distinct coauthors and a collaboration score of 70/100 — with middle-author, non-corresponding roles on both cataloged papers, consistent with team-based research. Reproducibility readiness is early-stage at a mean of 38/100 across two assessed papers (one medium, one low). Output, impact, and activity dimensions remain unscored for lack of data, and with no ORCID on record this profile is name-resolved from public bibliographic data.
Collaboration pattern
With 32 distinct coauthors recorded and a collaboration score of 70/100, Xinyi Mou works in sizeable teams. Recurring names include Xinnong Zhang, Jiayu Lin, Jia Wang, Yixu Huang, Yingqian Wu, Jingcong Liang, Shijun Lei, Jianing Shi, Guanying Li, Siyuan Wang, and Hanjia Lyu. On both cataloged 2026 papers the role is middle author without corresponding status, suggesting contribution within larger group efforts rather than sole leadership.
Research keywords
Highlights
State-grounded conditioning lifts agent direction accuracy from 26.5% to 83.5%
SocioVerse2: longitudinal social simulation under a human-AI co-evolutionary paradigm
Balanced two-field portfolio with a broad collaborative network
Field distribution splits 50/50 between LLM-based social simulation and LLM-based conversational agents, with 32 distinct coauthors recorded and middle-author roles on both cataloged papers — team-based work spanning two adjacent LLM-agents tracks.
Timeline
- 2026Released SocioVerse2 on arXiv: a longitudinal dynamic social simulation framework under a human-AI co-evolutionary paradigm.
- 2026Released the state-grounded conditioning wrapper for user-facing LLM agents on arXiv, reporting SGA improving from 26.5% to 83.5%.
- 2026Both papers entered the Evolve catalog, with reproducibility readiness rated medium (SocioVerse2) and low (state-grounded conditioning).
Papers in the Evolve catalog
Frequent co-authors
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 signals come from a limited snapshot: only two Evolve-reviewed arXiv preprints (both 2026) underpin the field distribution and reproducibility scores.
- Output, impact, and activity dimensions are unscored because publication counts, citation data, and multi-year productivity records were not available.
- No ORCID or country record was provided; the profile is name-resolved from public bibliographic data, and identity should be treated as provisional.
- The frequent-coauthor list contains an entry identical to the profile name (Xinyi Mou), likely a name-resolution artifact; it was excluded from collaboration analysis.