Benjamin Gruenbaum

Mayo Clinic in Florida · United States

Unclaimed · AI-structured profileOpenAlex

First-author builder of Era by Eon: synthetic enterprise data with exact ground truth for benchmarking LLM agents.

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-16) 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 underlying data does not support an output estimate. This axis would measure total and per-year publication counts; the visible record is limited to three 2026 arXiv preprints and a recent-three-year works count of 4.

Impact—

Not yet scored: citation data is too thin, with the snapshot reporting h-index 0 and i10-index 0. This axis would measure citation counts, citation velocity, and downstream influence once these 2026 preprints accumulate citations.

Focus0/100

The focus score is 0 because the field distribution splits evenly three ways (0.33 each): LLM agent benchmarking, LLM agent evaluation, and data systems / synthetic data generation for agentic AI. Numerically dispersed but thematically one program — generated enterprise data with exact ground truth for testing agents. OpenAlex topics (Topic Modeling 3, AI Applications 2, NLP Techniques 1) fit this line; two anesthesia-related topics likely reflect name homonymy.

Collaboration39/100

Collaboration scores 39, computed from 6 distinct coauthors across the record. All three cataloged 2026 preprints list Gruenbaum as first author with no corresponding-author flag, and the frequent-coauthor list — Doron Porat, Assaf Natanzon, Roy Zavida, Chen Dinachi, Or Itzahary, Omer Niv — indicates one stable in-house team rather than a broad external network.

Activity47/100

Activity scores 47. The counts-by-year series shows 4 works in 2026, and the recent-three-year works count is also 4 — all cataloged output is concentrated in 2026 with no earlier years in the snapshot, consistent with a newly started publication record.

Reproducibility43/100

Reproducibility scores 43, the mean across three reviewed papers. Evolve rates the two synthesis and benchmark-construction papers as medium reproducibility readiness and the original Era by Eon benchmark paper as low, citing unreleased code; recurring issues are single-seed evaluation, circular reference-free evaluation, and internal count errors.

Research summary

Benjamin Gruenbaum, affiliated with Mayo Clinic in Florida (US), is an early-stage researcher whose cataloged output concentrates on one applied problem: making large language model agents testable against enterprise data. All three works in the Evolve catalog are 2026 arXiv preprints with Gruenbaum as first author: the Era by Eon benchmark (a seed-generated fictional enterprise estate with exact ground truth for agent evaluation), a companion paper on deterministic synthesis and reference-free evaluation of multi-system business data, and an extension adding eight deterministic hidden-knowledge questions. The three preprints map onto the OpenAlex field distribution in equal thirds (0.33 each): LLM agent benchmarking, LLM agent evaluation, and data systems / synthetic data generation for agentic AI. Evolve's review verdicts credit the line of work with elegant, deterministic benchmark construction and honest limitation statements, while flagging single-seed evaluation, circular evaluation design, internal count errors, and unreleased code; the reproducibility mean across the three reviewed papers is 43/100 (readiness: medium, medium, low). The citation footprint is not yet established (h-index 0, i10-index 0), and all cataloged output appears within a compact six-person collaboration cluster.

Collaboration pattern

Gruenbaum's cataloged work is a compact-team effort: 6 distinct coauthors, all recurring — Doron Porat, Assaf Natanzon, Roy Zavida, Chen Dinachi, Or Itzahary, and Omer Niv. All three 2026 preprints list him as first author with no corresponding-author flag, suggesting he leads the writing while authorship is shared within one stable group rather than spread across a wide external network.

Research keywords

LLM agentsAgent benchmarkingHidden knowledgeSynthetic enterprise dataDeterministic data generationExact ground truthReference-free evaluationEntity graphAdversarial detectorQuestion-conditioned generationTool use evaluationEnterprise systems of record

Highlights

Extending Era by Eon with hidden-knowledge questions
Adds 8 deterministic hidden-knowledge questions to the Era by Eon enterprise-agent benchmark. Evolve's verdict praises an elegant construct and honest limits, but notes that a single-seed evaluation and internal count errors weaken the claims. Reproducibility readiness: medium.
Deterministic synthetic enterprise data with reference-free evaluation
Proposes deterministic synthesis of multi-system business data plus a reference-free evaluation pipeline built on an entity graph, adversarial detector, and question-conditioned generation. Evolve calls the engineering sound but flags a circular evaluation and missing external validation.
A generated enterprise estate with exact ground truth
The Era by Eon benchmark seed-generates fictional companies with exact answer keys for benchmarking LLM agents on systems of record and tool use. Evolve finds the validation sound, but the headline evaluation rests on one small estate and code is unreleased. Reproducibility readiness: low.

Timeline

2026Publishes 'The Era by Eon Benchmark: A Generated Enterprise Estate with Exact Ground Truth for Benchmarking LLM Agents' as first author on arXiv.
2026Publishes 'Generating a Consistent Enterprise' on deterministic synthesis and reference-free evaluation of multi-system business data, again as first author.
2026Extends the line of work with 'Era by Eon: Benchmarking Enterprise Agents on Hidden Knowledge', adding 8 deterministic hidden-knowledge questions.

Papers in the Evolve catalog

2026-09-24first author
Era by Eon: Benchmarking Enterprise Agents on Hidden Knowledge
2026-09-10first author
Generating a Consistent Enterprise: Synthesis and Reference-Free Evaluation of Multi-System Business Data
2026-09-09first author
The Era by Eon Benchmark: A Generated Enterprise Estate with Exact Ground Truth for Benchmarking LLM Agents

Frequent co-authors

Doron Porat×3 Assaf Natanzon×3 Roy Zavida×3 Chen Dinachi×3 Or Itzahary×3 Omer Niv×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.

Bibliometrics are derived from OpenAlex only; no ORCID is linked, so this profile is name-resolved from public bibliographic data and may be affected by name homonymy (OpenAlex lists two anesthesia-related topics absent from the cataloged preprints).
Output and impact dimensions are unscored: the snapshot offers no citation counts beyond h-index 0 and i10-index 0.
All cataloged works are 2026 arXiv preprints; no peer-reviewed venues, grants, or awards are in scope.
The reproducibility score (mean 43/100) reflects Evolve's review of 3 papers, not an independent replication study.