LLM agent evaluation and environment design researcher at Tencent; co-author of Env-Rethink and ExplorationBench
2
papers here
0
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
0
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-09-28) 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 scored: this snapshot provides no complete publication count or venue record (h-index and i10-index are recorded as 0, indicating incomplete author coverage). This axis would measure cumulative output volume and its distribution across venues and years.
Impact—
Not scored: no usable citation data exists for this author — h-index and i10-index are both recorded as 0, which here reflects missing citation coverage rather than measured absence of impact. This axis would aggregate citations, citation trajectory, and influential works.
Focus—
Not scored: the OpenAlex topic labels (Topic Modeling, 5 works; Multimodal Machine Learning, 3) hint at breadth, but the focus dimension is not yet computed. It would measure how concentrated output is across topics over time versus spread.
Collaboration66/100
The collaboration axis scores 66/100. The OpenAlex record lists 27 distinct coauthors, with 12 recurring collaborators including Weihuang Zheng, Ming Zhang, Zhenghao Xiang, and Shihan Dou. Both cataloged 2026 papers place this author in middle and last positions without corresponding authorship, suggesting work embedded in a larger LLM-agent team at Tencent rather than independently led projects.
Activity75/100
The activity axis scores 75/100. OpenAlex records 12 works in the most recent three-year window, and countsByYear attributes all 12 to 2026, indicating a concentrated recent burst rather than a long accumulation curve. Both cataloged papers are 2026 arXiv preprints on LLM agents, consistent with an active, current research phase. Longer-run activity levels cannot be read from this snapshot.
Reproducibility—
Not scored: no aggregate reproducibility assessment exists yet. It would roll up Evolve's repro-readiness ratings across reviewed papers; the only two assessments so far rate Env-Rethink medium and ExplorationBench low, too few for a reliable score.
Research summary
Maxm Pan is an AI researcher at Tencent (China) whose profiled work centers on large language model (LLM) agents — how to evaluate them and how to build and evolve the environments they act in. This is an early-stage profile, name-resolved from an OpenAlex bibliographic record. OpenAlex lists 12 works in the most recent three-year window, all attributed to 2026, alongside 27 distinct coauthors; topic labels include Topic Modeling, Multimodal Machine Learning Applications, Scientific Computing and Data Management, Explainable AI, and Software Engineering, suggesting a text- and data-oriented background now applied to agent systems. Evolve has reviewed two 2026 arXiv papers involving this author. In "Breaking the Environment Wall" (Env-Rethink, middle author), a 27B verifier lifted noisy-environment agent pass rates by 13.3 percentage points across 9 models and evolved harder tasks, though reviewers judged recursive self-improvement unproven. In "ExplorationBench" (last author), a benchmark measures AI systems' exploration in verifiable alien worlds with exact grading; its design is considered strong, but n=3 trajectories and single-run controls limit statistical strength. Citation-level impact cannot yet be assessed: h-index and i10-index are recorded as 0, which most plausibly reflects incomplete author-level indexing rather than measured absence of impact.
Collaboration pattern
The collaboration axis scores 66/100. The OpenAlex record lists 27 distinct coauthors, with 12 recurring collaborators including Weihuang Zheng, Ming Zhang, Zhenghao Xiang, Peizhong Gao, Yujiong Shen, and Shihan Dou. In both reviewed papers this author holds middle or last authorship without corresponding status, consistent with contributing inside a larger Tencent LLM-agent team rather than leading independent projects.
Research keywords
Highlights
Env-Rethink: a 27B verifier lifts noisy-environment agent pass rates by 13.3pp
ExplorationBench: grading AI exploration in verifiable alien worlds
Timeline
- 2026Co-authored Env-Rethink (middle author, arXiv): a 27B verifier lifts noisy-environment agent pass rates by 13.3pp across 9 models and evolves harder tasks.
- 2026Co-authored ExplorationBench (last author, arXiv): a benchmark measuring AI systems' exploration in verifiable alien worlds with exact grading.
- 2026OpenAlex attributes 12 works to 2026 with 27 distinct coauthors, the most active recorded year in this snapshot.
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.
- Bibliometrics are name-resolved from an OpenAlex record; h-index and i10-index are recorded as 0, indicating incomplete author-level citation coverage rather than measured zero impact.
- Only two papers in this snapshot are Evolve-reviewed, both 2026 arXiv preprints; peer-reviewed venue status is not represented.
- No ORCID is provided, so identity resolution relies solely on the OpenAlex profile for 'Maxm Pan' at Tencent (China).
- Citation counts, full publication history, and venue list are unavailable in this snapshot; dimensional scores other than collaboration and activity cannot be computed.