NLP researcher at Concordia University: LLM contextual causality and creativity metrics for automated research
2
papers here
2
reviewed by Evolve
2
first-author
0
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: no publication-count, venue, or output-volume metrics are available in the current data. This axis would measure the volume and consistency of published research across venues and years, none of which can be assessed here.
Impact—
Not yet scored: no citation counts, h-index, or influence indicators are available. This axis would measure how widely the work is cited and taken up; the two 2026 arXiv preprints are too recent for such signals in this snapshot.
Focus0/100
The composite focus score reads 0, but the underlying field distribution shows a 50/50 split between two closely adjacent areas — natural language processing and NLP for LLM agents / automated research. Read literally, the score suggests dispersion; read against the raw field labels, the agenda is tightly concentrated on LLM reasoning and evaluation. Treat the 0 as a labeling artifact until more papers widen or confirm the picture.
Collaboration36/100
Collaboration scores 36, based on 5 distinct coauthors in the current snapshot. The recurring circle is compact — Chengming Hu, Mengzhuo Chen, Pengyi Liao, Yihan Huang, and Jun Yan. Both catalog papers are first-authored, pointing to a leading role within a small, stable team rather than a large distributed collaboration network.
Activity—
Not yet scored: the snapshot contains no yearly publication counts, so publishing cadence and sustained activity cannot be assessed. This axis would measure output rhythm over time; only two papers dated 2026 are currently observable.
Reproducibility38/100
Reproducibility averages 38 across the 2 papers with readiness ratings. The contextual-causality survey is rated low, reflecting narrative synthesis without a systematic search protocol, while the creativity-metrics study is rated medium; its headline correlation rests on n=6 with p=0.063, a small-sample limitation the Evolve review explicitly flags.
Research summary
Yiheng Zhao is a natural language processing researcher at Concordia University whose recent work centers on how large language models handle contextual causality and how the creativity of LLM-driven automated research can be measured. The Evolve catalog lists two 2026 arXiv papers, both first-authored: a survey of contextual causality with LLMs, which organizes the literature around semantic, intervention, and counterfactual perspectives, and a study proposing interpretable valueness and novelty metrics for frontier LLM research agents. The bibliometric field distribution splits evenly — 50% core natural language processing, 50% NLP for LLM agents / automated research — indicating a tightly concentrated early agenda rather than scattered interests. Collaboration is small-scale: 5 distinct coauthors, with Chengming Hu, Mengzhuo Chen, Pengyi Liao, Yihan Huang, and Jun Yan recurring most often; the collaboration dimension scores 36. Reproducibility averages 38 across the 2 papers, matching one survey rated low in reproducibility readiness and one empirical study rated medium, whose headline correlation rests on n=6 with p=0.063 — a small-sample limitation the Evolve review explicitly flags. Output, impact, and activity dimensions lack sufficient data and are left unscored. This is an early-stage profile built from a small, arXiv-only evidence base.
Collaboration pattern
Collaboration runs through a compact five-person circle: Chengming Hu, Mengzhuo Chen, Pengyi Liao, Yihan Huang, and Jun Yan are the most frequent coauthors, matching the 5 distinct coauthors in the bibliometric snapshot. The collaboration dimension scores 36, consistent with an early-stage network. Both catalog papers are first-authored but not corresponding-author work, and the same coauthors recur, pointing to a stable small project team rather than ad-hoc pairings.
Research keywords
Highlights
Surveying contextual causality with LLMs via a semantic–intervention–counterfactual taxonomy
Interpretable valueness and novelty metrics for frontier LLM research agents
Timeline
- 2026Publishes first-authored arXiv survey "Contextual Causality with Large Language Models: A Survey", introducing a semantic/intervention/counterfactual taxonomy for LLM contextual causality.
- 2026Publishes first-authored arXiv study "Measuring the Creativity of Frontier LLMs in Automated Research", proposing interpretable valueness and novelty metrics for LLM research agents.
- 2026Both 2026 outputs map onto the two adjacent halves of the field distribution — core NLP and LLM agents/automated research — consolidating a two-pronged early research agenda.
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 dimensions are computed from OpenAlex-derived data only; citation, venue-tier, and publication-volume metrics were unavailable, so output, impact, and activity are left unscored.
- Evidence base is limited to 2 arXiv preprints from 2026 reviewed in the Evolve catalog — an early-stage profile that may not reflect the full publication record.
- No ORCID or country is attached; "Yiheng Zhao" is a moderately common name, so this profile is name-resolved from public bibliographic data and may merge or split identities.
- Dimension scores use Evolve's 0-100 composite scale; the focus score of 0 reflects a labeling artifact of two adjacent field names rather than measured dispersion.