Configuration

LLM Ensemble (AssayBench)

Configuration of LLM Ensemble: ensemble over several model outputs, discovered with AlphaEvolve on the training split.

Overview

An explanatory profile has not yet been reviewed for this record.

Consult the linked sources for architecture or protocol details. Missing evidence is not evidence of a missing capability.

Evaluations and results

3 evaluations · 9 results. Different protocols are not a single leaderboard.

Protocol: AssayBench: rank 100 candidate genes for a described CRISPR screen

Sorted by Adjusted NDCG at 100 (post-cutoff cohort; mean over benchmark entries) (higher is better). The best value in each column is highlighted. Decimals are rounded for display; each value links to the printed value and its source.

Tested configuration	Protocol and dataset	Adjusted NDCG at 100 (post-cutoff cohort; mean over benchmark entries)	Adjusted NDCG at 100 (test cohort; mean over benchmark entries)	Adjusted NDCG at 100 (validation cohort; mean over benchmark entries)	Directional false discovery rate at 100 (post-cutoff cohort; mean over benchmark entries)	Directional false discovery rate at 100 (test cohort; mean over benchmark entries)	Directional false discovery rate at 100 (validation cohort; mean over benchmark entries)	Precision at 100 (post-cutoff cohort; mean over benchmark entries; denominator min(100, positive-relevance genes))	Precision at 100 (test cohort; mean over benchmark entries; denominator min(100, positive-relevance genes))	Precision at 100 (validation cohort; mean over benchmark entries; denominator min(100, positive-relevance genes))
LLM Ensemble (AssayBench)Evaluation	AssayBench post-cutoff cohort of human CRISPR screens	0.109	Not reported	Not reported	NA	Not reported	Not reported	0.248	Not reported	Not reported
LLM Ensemble (AssayBench)Evaluation	AssayBench test cohort of human CRISPR screens	Not reported	0.163	Not reported	Not reported	0.0178	Not reported	Not reported	0.22	Not reported
LLM Ensemble (AssayBench)Evaluation	AssayBench validation cohort of human CRISPR screens	Not reported	Not reported	0.186	Not reported	Not reported	0.0145	Not reported	Not reported	0.324

All 9 result rows with coverage, uncertainty and sources

Filter evaluations

Tested model or methodProtocolDatasetEvaluation setupMetricEvidence origin

Applied filters: All linked evaluations

Exact evaluated configurations and original reported results
Tested configuration	Protocol and dataset	Finding	Evidence and details
Configuration: LLM Ensemble (AssayBench)	Protocol: AssayBench: rank 100 candidate genes for a described CRISPR screen Dataset: AssayBench post-cutoff cohort of human CRISPR screens	0.109 Adjusted NDCG at 100 (post-cutoff cohort; mean over benchmark entries) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source LLM Ensemble on the AssayBench post-cutoff cohort tgtval-20261009-protocol-debrouwer2026-screen-gene-ranking Aggregation: Not reported AssayBench: An Assay-Level Virtual Cell Benchmark for LLMs and Agents · Table 3, row 'LLM Ensemble' cohort 'LaTest', column 'AnDCG@100'
Configuration: LLM Ensemble (AssayBench)	Protocol: AssayBench: rank 100 candidate genes for a described CRISPR screen Dataset: AssayBench post-cutoff cohort of human CRISPR screens	NA Directional false discovery rate at 100 (post-cutoff cohort; mean over benchmark entries) fraction · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source LLM Ensemble on the AssayBench post-cutoff cohort tgtval-20261009-protocol-debrouwer2026-screen-gene-ranking Aggregation: Not reported AssayBench: An Assay-Level Virtual Cell Benchmark for LLMs and Agents · Table 3, row 'LLM Ensemble' cohort 'LaTest', column 'dFDR@100'
Configuration: LLM Ensemble (AssayBench)	Protocol: AssayBench: rank 100 candidate genes for a described CRISPR screen Dataset: AssayBench post-cutoff cohort of human CRISPR screens	0.248 Precision at 100 (post-cutoff cohort; mean over benchmark entries; denominator min(100, positive-relevance genes)) fraction · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source LLM Ensemble on the AssayBench post-cutoff cohort tgtval-20261009-protocol-debrouwer2026-screen-gene-ranking Aggregation: Not reported AssayBench: An Assay-Level Virtual Cell Benchmark for LLMs and Agents · Table 3, row 'LLM Ensemble' cohort 'LaTest', column 'Precision@100'
Configuration: LLM Ensemble (AssayBench)	Protocol: AssayBench: rank 100 candidate genes for a described CRISPR screen Dataset: AssayBench test cohort of human CRISPR screens	0.163 Adjusted NDCG at 100 (test cohort; mean over benchmark entries) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source LLM Ensemble on the AssayBench test cohort tgtval-20261009-protocol-debrouwer2026-screen-gene-ranking Aggregation: Not reported AssayBench: An Assay-Level Virtual Cell Benchmark for LLMs and Agents · Table 3, row 'LLM Ensemble' cohort 'test', column 'AnDCG@100'
Configuration: LLM Ensemble (AssayBench)	Protocol: AssayBench: rank 100 candidate genes for a described CRISPR screen Dataset: AssayBench test cohort of human CRISPR screens	0.0178 Directional false discovery rate at 100 (test cohort; mean over benchmark entries) fraction · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source LLM Ensemble on the AssayBench test cohort tgtval-20261009-protocol-debrouwer2026-screen-gene-ranking Aggregation: Not reported AssayBench: An Assay-Level Virtual Cell Benchmark for LLMs and Agents · Table 3, row 'LLM Ensemble' cohort 'test', column 'dFDR@100'
Configuration: LLM Ensemble (AssayBench)	Protocol: AssayBench: rank 100 candidate genes for a described CRISPR screen Dataset: AssayBench test cohort of human CRISPR screens	0.22 Precision at 100 (test cohort; mean over benchmark entries; denominator min(100, positive-relevance genes)) fraction · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source LLM Ensemble on the AssayBench test cohort tgtval-20261009-protocol-debrouwer2026-screen-gene-ranking Aggregation: Not reported AssayBench: An Assay-Level Virtual Cell Benchmark for LLMs and Agents · Table 3, row 'LLM Ensemble' cohort 'test', column 'Precision@100'
Configuration: LLM Ensemble (AssayBench)	Protocol: AssayBench: rank 100 candidate genes for a described CRISPR screen Dataset: AssayBench validation cohort of human CRISPR screens	0.186 Adjusted NDCG at 100 (validation cohort; mean over benchmark entries) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source LLM Ensemble on the AssayBench validation cohort tgtval-20261009-protocol-debrouwer2026-screen-gene-ranking Aggregation: Not reported AssayBench: An Assay-Level Virtual Cell Benchmark for LLMs and Agents · Table 3, row 'LLM Ensemble' cohort 'val', column 'AnDCG@100'
Configuration: LLM Ensemble (AssayBench)	Protocol: AssayBench: rank 100 candidate genes for a described CRISPR screen Dataset: AssayBench validation cohort of human CRISPR screens	0.0145 Directional false discovery rate at 100 (validation cohort; mean over benchmark entries) fraction · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source LLM Ensemble on the AssayBench validation cohort tgtval-20261009-protocol-debrouwer2026-screen-gene-ranking Aggregation: Not reported AssayBench: An Assay-Level Virtual Cell Benchmark for LLMs and Agents · Table 3, row 'LLM Ensemble' cohort 'val', column 'dFDR@100'
Configuration: LLM Ensemble (AssayBench)	Protocol: AssayBench: rank 100 candidate genes for a described CRISPR screen Dataset: AssayBench validation cohort of human CRISPR screens	0.324 Precision at 100 (validation cohort; mean over benchmark entries; denominator min(100, positive-relevance genes)) fraction · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source LLM Ensemble on the AssayBench validation cohort tgtval-20261009-protocol-debrouwer2026-screen-gene-ranking Aggregation: Not reported AssayBench: An Assay-Level Virtual Cell Benchmark for LLMs and Agents · Table 3, row 'LLM Ensemble' cohort 'val', column 'Precision@100'

Source checking is not independent reproduction. Release 2026-11-02-a7f03c9e55a5.

Use this configuration

How it works, versions and access

Additional methodological details, versions and access terms have not yet been extracted into a structured profile for this configuration. Consult the linked sources and evaluation records below.

Evidence

Source checking verifies the cited claim or transcription. It does not establish independent reproduction.

Evidence table

Inspect claims, sources and review details

Trace each statement to its source and review. A context-only reference supports the record generally; it does not verify an individual field. Source checking does not reproduce an experiment.

One row per statement and cited source. Multiple citations are not independent evaluations. Shared locators are labelled explicitly.

0 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-11-02-a7f03c9e55a5
Property and statement	Original source and location	Review and provenance

No evidence rows match these filters. Choose another scope or clear the search.

How to read the evidence

Sources and history

Release 2026-11-02-a7f03c9e55a5 · Record review: source checked

1 source record and release history

AssayBench: An Assay-Level Virtual Cell Benchmark for LLMs and Agents · Original source · arXiv:2605.10876 version 1, posted 2026-05-11; not peer reviewed

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: tgtval-20261009-config-debrouwer2026-llm-ensemble

areas: cells-tissues
contexts: research
method types: foundation_model
reported name: LLM Ensemble
foundation model eligible: true
source locator: Tables 2-3 row labels; section 4
missing metadata: version: reason: unreported; note: No release or commit is printed for this configuration
parameters: ensemble over several model outputs, discovered with AlphaEvolve on the training split
provenance note: Configuration as run in the cited comparison.

Related records

configuration of: LLM Ensemble
system: LLM Ensemble on the AssayBench post-cutoff cohort
system: LLM Ensemble on the AssayBench test cohort
system: LLM Ensemble on the AssayBench validation cohort

Suggest a correction