Result

0.199 precision-at-100

debrouwer2026-gemini-3-flash-gepa-test precision-at-100

Tested configuration: Gemini 3 Flash (GEPA) (AssayBench)
Protocol: AssayBench: rank 100 candidate genes for a described CRISPR screen
Dataset: AssayBench test cohort of human CRISPR screens
Procedure: tgtval-20261009-protocol-debrouwer2026-screen-gene-ranking
Evaluation: Gemini 3 Flash (GEPA) on the AssayBench test cohort
Coverage: Not reported scored / Not reported eligible
Uncertainty: Not reported by the source
Evidence: Author-reported evaluation · source checkedAssayBench: An Assay-Level Virtual Cell Benchmark for LLMs and Agents · Table 3, row 'Gemini 3 Flash (GEPA)' cohort 'test', column 'Precision@100'

A source-checked result verifies the numerical transcription, not every model or protocol detail. Evaluation metadata: source checked. Source checked does not mean independently reproduced.

Reproduction

Split: Temporal split, test cohort
Adaptation: prompts optimised with GEPA on the training split
Scoring implementation: AnDCG@100, Precision@100 and dFDR@100 as defined in section 3

No execution recipe has been verified for this exact configuration and evaluation. A benchmark's general instructions may use different inputs, splits or model settings.

Reproducing this published result requires matching its model configuration, data, split and scorer. Source checking or a successful smoke test does not establish score reproduction.

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.

1 evidence row matching the loaded filters

Claims, original sources and review scope · Release 2026-10-10-6e93f504adfc
Property and statement	Original source and location	Review and provenance
Reported result 0.1987 Individual claims	AssayBench: An Assay-Level Virtual Cell Benchmark for LLMs and Agents Original source ↗ Table 3, row 'Gemini 3 Flash (GEPA)' cohort 'test', column 'Precision@100' Version: arXiv:2605.10876 version 1, posted 2026-05-11; not peer reviewed Retrieved: 2026-10-09T20:56:07Z	source checked ["source-hash-verification","pdf-text-parse","independent-cell-check"] · 2026-10-09T21:22:54Z author reported Audit details Extracted by deterministic parse of the PDF text layer (pdftotext -layout) of Table 3, asserting the column header, 45 rows, 15 systems, the three cohorts per system and that dFDR is NA only for the post-cutoff cohort. Independent review 2026-10-09: value, metric, unit, direction, locator and configuration, protocol and dataset identity match the source. The qualifier carries the AssayBench denominator so these values are never compared with precision over a fixed 100 predictions. Field: `attributes.printed_value` Source artifact SHA-256: `805b402c28e0daa186415af202e04df56bf0cd7c7b1f872a6db170ee3c6c623d` Hash scope: pdftotext -layout text layer, parsed by extract/extract_target_validation.py Inspected artifact Extraction artifact SHA-256: `805b402c28e0daa186415af202e04df56bf0cd7c7b1f872a6db170ee3c6c623d` Extraction artifact

All evidence (JSONL) (gzip)How to read the evidence

Sources and history

Release 2026-10-10-6e93f504adfc · Record review: source checked

1 source records 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-result-debrouwer2026-gemini-3-flash-gepa-test-precision-at-100

metric: precision-at-100
metric qualifier: test cohort; mean over benchmark entries; denominator min(100, positive-relevance genes)
metric direction: higher
unit: fraction
printed value: 0.1987
numeric value: 0.1987
source locator: Table 3, row 'Gemini 3 Flash (GEPA)' cohort 'test', column 'Precision@100'
review: method: source-hash-verification; pdf-text-parse; independent-cell-check; method note: Re-downloaded the PDF and matched its SHA-256. Converted with pdftotext -layout and parsed Table 3 with a separate script written for this review, asserting the header and 48 rows. Checked printed and numeric value, metric, qualifier, unit, direction, locator, and the linked configuration, protocol and cohort dataset.; reviewer: claude; reviewer note: Separate Claude review agent, independent of the extractor; no human review claimed; reviewed at: 2026-10-09T21:22:54Z; artifact sha256: 805b402c28e0daa186415af202e04df56bf0cd7c7b1f872a6db170ee3c6c623d; retrieval url: https://arxiv.org/pdf/2605.10876v1; note: Extracted by deterministic parse of the PDF text layer (pdftotext -layout) of Table 3, asserting the column header, 45 rows, 15 systems, the three cohorts per system and that dFDR is NA only for the post-cutoff cohort. Independent review 2026-10-09: value, metric, unit, direction, locator and configuration, protocol and dataset identity match the source. The qualifier carries the AssayBench denominator so these values are never compared with precision over a fixed 100 predictions.
missing metadata: uncertainty: reason: unreported
scope note: Precision@100 denominator is min(100, positive-relevance genes) (section 3.2).

Related records

evaluation: Gemini 3 Flash (GEPA) on the AssayBench test cohort

Suggest a correction