LLM agent evaluation researcher; first author of tool-augmented deductive reasoning work in multi-agent games (Clue)
1
papers here
1
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-09-21) 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—
No output metrics (publication count and output volume) are available in this snapshot yet; this axis would measure the volume and consistency of publication output over time. The catalog currently holds one 2026 arXiv first-author paper.
Impact—
Impact metrics like citation counts are not yet available; this axis would measure how widely the work is cited and taken up. With a single 2026 arXiv paper in the catalog, citation-based impact cannot be scored.
Focus100/100
A perfect focus score of 100 reflects a single-field distribution: all indexed work sits in artificial intelligence (LLM agent evaluation). The 2026 paper's keywords — LLM agents, tool augmentation, deductive reasoning, belief state tracking — confirm a concentrated research identity, though one paper is a narrow base for judging long-term focus.
Collaboration14/100
Scores 14 with exactly one distinct coauthor in the snapshot (Autumn Toney-Wails). The 2026 Clue paper is a first-author, two-person collaboration, so the current footprint is a small, tightly focused pair rather than a broad network; the low score reflects thin co-authorship data from a single indexed paper, and the pattern may change as more work is published.
Activity—
No activity metrics (recent publication cadence or engagement) are available yet; this axis would measure ongoing research output and community participation. The snapshot records one 2026 arXiv preprint, too little to characterize activity.
Reproducibility55/100
A mean reproducibility-readiness score of 55, computed from one reviewed paper (reproPapers: 1), sits at mid-range. Evolve rated the 2026 Clue study medium: near-perfect deduction results rest on only six games per condition, and the tool itself does much of the deductive work, so replication would need broader game sampling.
Research summary
Rebecca Ansell is a researcher whose visible work sits at the intersection of large language model agents and deductive reasoning evaluation; this is an early-stage profile built from a single indexed paper. The Evolve catalog holds one entry, the 2026 arXiv preprint 'Clueing up LLMs with Tool-Augmented Deductive Reasoning', on which she is first author, co-written with Autumn Toney-Wails. The study equips LLM agents with a possibility-matrix tool for belief state tracking in the Clue board game and reports near-perfect deductive performance in multi-agent game settings. Evolve's review adds an important qualification: the possibility-matrix tool largely solves the game itself, so the headline results speak to tool design as much as to the underlying model's reasoning, and the experimental base is small at six games per condition. The paper is rated medium on reproducibility readiness, producing the mean reproducibility score of 55 in this profile. Bibliometrically the snapshot is deliberately narrow: one indexed paper, one distinct co-author, and a single field — artificial intelligence (LLM agent evaluation) — which yields a perfect focus score of 100 and a low collaboration score of 14. Output, impact, and activity dimensions are not yet scored because the underlying metrics are missing. As a first-author contribution to LLM agent evaluation, the work signals a research identity centred on tool augmentation, deductive reasoning, and multi-agent game benchmarks, but the profile should be read as an early-stage snapshot likely to evolve with further publications.
Collaboration pattern
The current record shows a deliberately small collaboration footprint: exactly one distinct co-author, Autumn Toney-Wails, who co-authored the 2026 arXiv first-author paper on tool-augmented deductive reasoning. This yields a low collaboration score of 14, but it is based on a single indexed paper, so it signals an early, focused two-person working pattern rather than an established collaborative network.
Research keywords
Highlights
Possibility-matrix tool drives near-perfect LLM deduction in Clue
Evolve verdict: the tool largely solves the game itself
Evaluation rests on six games per condition
Timeline
- 2026First-author arXiv preprint 'Clueing up LLMs with Tool-Augmented Deductive Reasoning' posted, in the field of LLM agent evaluation.
- 2026Co-authored the paper with Autumn Toney-Wails — the only co-author recorded in the current snapshot.
- 2026Evolve review logged with medium reproducibility readiness (mean 55), noting the tool largely solves the game itself and results rest on six games per condition.
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 scores are computed from OpenAlex-sourced data only; other indexes may list additional papers or co-authors not reflected here.
- The profile rests on a single reviewed paper (2026 arXiv), so all dimensions are early-stage and may shift substantially.
- No institution, country, or ORCID is attached; the profile is name-resolved from public bibliographic data, and 'Rebecca Ansell' may be an ambiguous name.
- Output, impact, and activity axes are unscored because the underlying citation and volume metrics were missing from this snapshot.