Source

Zero-shot evaluation reveals limitations of single-cell foundation models (Kedzierska et al. 2025, Genome Biology 26:101)

Genome Biology · 2025 · Version of record, Genome Biology 26:101, published 2025-04-18 · Peer-reviewed

What this source contributed

Records in this release that cite this source. A record can cite several sources.

1 method
8 configurations
1 benchmark
3 protocols
7 dataset subsets
113 evaluations
778 results
4 evidence claims

Results

Values as the source printed them. Compare results only under compatible protocols, datasets and metrics.

Claims

Statements this source supports, each with where in the source it was checked.

Immune (330k): overlap with scGPT pretraining (Kedzierska et al. 2025): Pretraining overlap:scGPT The Immune (Cross-tissue Immune Cell Atlas) dataset was included in the scGPT pretraining data (CELLxGENE May 2023 census). (Additional file 1, sheet 'TableS2' cell H12 (column 'scGPT', row 'Immune (330k)'); page 9 'Datasets')
Pancreas (8k) (Kedzierska et al. 2025): description: Description Row labelled 'Pancreas (8k)' in the score sheets: the Pancreas subset without GEO GSE84133, which was in Geneformer pretraining. The Methods (page 9) say both the complete Pancreas dataset and its non-overlapping subset were scored 'in Additional file 1: Tables S1 and S5', which are the two score sheets, and this is their only extra Pancreas row. The workbook itself does not name the subset. (Article page 9 'Datasets' paragraph 3; Additional file 1 sheets TableS6 and TableS7 row 'Pancreas (8k)')
Pancreas (16k): overlap with Geneformer pretraining (Kedzierska et al. 2025): Pretraining overlap:Geneformer The Pancreas study GSE84133 (Baron et al.) was included in Geneformer pretraining; the other five Pancreas studies were not. (Additional file 1, sheet 'TableS2' cell G6 (column 'Geneformer', row 'GEO - GSE84133'); page 9 'Datasets')
Tabula Sapiens (483k): overlap with scGPT pretraining (Kedzierska et al. 2025): Pretraining overlap:scGPT Tabula Sapiens was included in the scGPT pretraining data (CELLxGENE May 2023 census). (Additional file 1, sheet 'TableS2' cell H11 (column 'scGPT', row 'Tabula Sapiens (483k)'); page 9 'Datasets')

Models, benchmarks and other records

Methods (1)

Highly variable gene selection

Configurations (8)

Benchmarks (1)

Zero-shot cell embedding evaluation of single-cell foundation models (Kedzierska et al. 2025)

Protocols (3)

Dataset subsets (7)

Evaluations (113)

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.

28 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-10-23-5347695500ed
Property and statement	Original source and location	Review and provenance
Artifact SHA-256 8efcc5fcb2e35b9a04b1ee9e03581e20fffeb5521a4a4c748a6f9b34f7d3c47f Source metadata	Zero-shot evaluation reveals limitations of single-cell foundation models (Kedzierska et al. 2025, Genome Biology 26:101) Original source ↗ No field-specific location recorded Version: Version of record, Genome Biology 26:101, published 2025-04-18 Retrieved: 2026-10-14T12:01:07Z	catalogued No individual claim review recorded Audit details Field: `attributes.artifact_sha256` Source artifact SHA-256: `8efcc5fcb2e35b9a04b1ee9e03581e20fffeb5521a4a4c748a6f9b34f7d3c47f` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Artifact url https://genomebiology.biomedcentral.com/counter/pdf/10.1186/s13059-025-03574-x.pdf Source metadata	Zero-shot evaluation reveals limitations of single-cell foundation models (Kedzierska et al. 2025, Genome Biology 26:101) Original source ↗ No field-specific location recorded Version: Version of record, Genome Biology 26:101, published 2025-04-18 Retrieved: 2026-10-14T12:01:07Z	catalogued No individual claim review recorded Audit details Field: `attributes.artifact_url` Source artifact SHA-256: `8efcc5fcb2e35b9a04b1ee9e03581e20fffeb5521a4a4c748a6f9b34f7d3c47f` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Doi 10.1186/s13059-025-03574-x Source metadata	Zero-shot evaluation reveals limitations of single-cell foundation models (Kedzierska et al. 2025, Genome Biology 26:101) Original source ↗ No field-specific location recorded Version: Version of record, Genome Biology 26:101, published 2025-04-18 Retrieved: 2026-10-14T12:01:07Z	catalogued No individual claim review recorded Audit details Field: `attributes.doi` Source artifact SHA-256: `8efcc5fcb2e35b9a04b1ee9e03581e20fffeb5521a4a4c748a6f9b34f7d3c47f` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Licence CC-BY-4.0 Source metadata	Zero-shot evaluation reveals limitations of single-cell foundation models (Kedzierska et al. 2025, Genome Biology 26:101) Original source ↗ No field-specific location recorded Version: Version of record, Genome Biology 26:101, published 2025-04-18 Retrieved: 2026-10-14T12:01:07Z	catalogued No individual claim review recorded Audit details Field: `attributes.licence` Source artifact SHA-256: `8efcc5fcb2e35b9a04b1ee9e03581e20fffeb5521a4a4c748a6f9b34f7d3c47f` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Limitations 2 values The headline comparisons (Fig. 1B average BIO score, Fig. 1E average batch score, Fig. 2 reconstruction) are figures without printed values; the per-dataset scores are in Additional file 1. All five evaluation datasets were chosen by the authors; Tabula Sapiens and Immune were in scGPT pretraining and part of Pancreas in Geneformer pretraining (Additional file 1, sheet TableS2). Source metadata	Zero-shot evaluation reveals limitations of single-cell foundation models (Kedzierska et al. 2025, Genome Biology 26:101) Original source ↗ No field-specific location recorded Version: Version of record, Genome Biology 26:101, published 2025-04-18 Retrieved: 2026-10-14T12:01:07Z	catalogued No individual claim review recorded Audit details Field: `attributes.limitations` Source artifact SHA-256: `8efcc5fcb2e35b9a04b1ee9e03581e20fffeb5521a4a4c748a6f9b34f7d3c47f` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Media type application/pdf Source metadata	Zero-shot evaluation reveals limitations of single-cell foundation models (Kedzierska et al. 2025, Genome Biology 26:101) Original source ↗ No field-specific location recorded Version: Version of record, Genome Biology 26:101, published 2025-04-18 Retrieved: 2026-10-14T12:01:07Z	catalogued No individual claim review recorded Audit details Field: `attributes.media_type` Source artifact SHA-256: `8efcc5fcb2e35b9a04b1ee9e03581e20fffeb5521a4a4c748a6f9b34f7d3c47f` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Publication status peer_reviewed Source metadata	Zero-shot evaluation reveals limitations of single-cell foundation models (Kedzierska et al. 2025, Genome Biology 26:101) Original source ↗ No field-specific location recorded Version: Version of record, Genome Biology 26:101, published 2025-04-18 Retrieved: 2026-10-14T12:01:07Z	catalogued No individual claim review recorded Audit details Field: `attributes.publication_status` Source artifact SHA-256: `8efcc5fcb2e35b9a04b1ee9e03581e20fffeb5521a4a4c748a6f9b34f7d3c47f` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Retrieved at 2026-10-14T12:01:07Z Source metadata	Zero-shot evaluation reveals limitations of single-cell foundation models (Kedzierska et al. 2025, Genome Biology 26:101) Original source ↗ No field-specific location recorded Version: Version of record, Genome Biology 26:101, published 2025-04-18 Retrieved: 2026-10-14T12:01:07Z	catalogued No individual claim review recorded Audit details Field: `attributes.retrieved_at` Source artifact SHA-256: `8efcc5fcb2e35b9a04b1ee9e03581e20fffeb5521a4a4c748a6f9b34f7d3c47f` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Review: artifact sha256 8efcc5fcb2e35b9a04b1ee9e03581e20fffeb5521a4a4c748a6f9b34f7d3c47f Source metadata	Zero-shot evaluation reveals limitations of single-cell foundation models (Kedzierska et al. 2025, Genome Biology 26:101) Original source ↗ No field-specific location recorded Version: Version of record, Genome Biology 26:101, published 2025-04-18 Retrieved: 2026-10-14T12:01:07Z	catalogued No individual claim review recorded Audit details Field: `attributes.review.artifact_sha256` Source artifact SHA-256: `8efcc5fcb2e35b9a04b1ee9e03581e20fffeb5521a4a4c748a6f9b34f7d3c47f` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Review: date 2026-10-14 Source metadata	Zero-shot evaluation reveals limitations of single-cell foundation models (Kedzierska et al. 2025, Genome Biology 26:101) Original source ↗ No field-specific location recorded Version: Version of record, Genome Biology 26:101, published 2025-04-18 Retrieved: 2026-10-14T12:01:07Z	catalogued No individual claim review recorded Audit details Field: `attributes.review.date` Source artifact SHA-256: `8efcc5fcb2e35b9a04b1ee9e03581e20fffeb5521a4a4c748a6f9b34f7d3c47f` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact

How to read the evidence

Sources and history

Release 2026-10-23-5347695500ed · Record review: source checked

Publisher PDF of the article, retrieved and hashed for the 2026-10-14 genomic foundation model guide fixes (issue rewire-benchmark-data#129).

Release history

No supporting source is linked yet.

Read original source

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: gfm-guides-20261014-kedzierska2025-source-article

areas: single-cell
contexts: research
url: https://doi.org/10.1186/s13059-025-03574-x
artifact url: https://genomebiology.biomedcentral.com/counter/pdf/10.1186/s13059-025-03574-x.pdf
version: Version of record, Genome Biology 26:101, published 2025-04-18
retrieved at: 2026-10-14T12:01:07Z
artifact sha256: 8efcc5fcb2e35b9a04b1ee9e03581e20fffeb5521a4a4c748a6f9b34f7d3c47f
doi: 10.1186/s13059-025-03574-x
publication status: peer_reviewed
licence: CC-BY-4.0
media type: application/pdf
venue: Genome Biology
year: 2025
limitations: The headline comparisons (Fig. 1B average BIO score, Fig. 1E average batch score, Fig. 2 reconstruction) are figures without printed values; the per-dataset scores are in Additional file 1.; All five evaluation datasets were chosen by the authors; Tabula Sapiens and Immune were in scGPT pretraining and part of Pancreas in Geneformer pretraining (Additional file 1, sheet TableS2).
review: method: source-hash-verification; method note: A fresh download on 2026-10-14 returned a 3,038-byte bot-check HTML page instead of the PDF; the archived copy in artifacts/ hashes to the pinned SHA-256 and was used.; reviewer: claude; reviewer note: Claude (Opus 5.5) research agent, independent of the extractor; no human review claimed; date: 2026-10-14; artifact sha256: 8efcc5fcb2e35b9a04b1ee9e03581e20fffeb5521a4a4c748a6f9b34f7d3c47f; retrieval url: https://genomebiology.biomedcentral.com/counter/pdf/10.1186/s13059-025-03574-x.pdf; note: Independent review 2026-10-14.

Related records

Suggest a correction

Model or configuration	Metric	Value	Dataset	Evaluation
Geneformer (Kedzierska et al. 2025)	Rank (overall average ranking across datasets, based on mean biological-conservation and batch-mixing scores (1 = best); full embedding dimensions)	8 unitless	Evaluated datasets, aggregated (Kedzierska et al. 2025)	Geneformer cell embedding on evaluated datasets (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Rank (rank of the median ARI score across datasets among the eight representations (1 = best); full embedding dimensions)	7 unitless	Evaluated datasets, aggregated (Kedzierska et al. 2025)	Geneformer cell embedding on evaluated datasets (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Rank (rank of the median ASW score across datasets among the eight representations (1 = best); full embedding dimensions)	8 unitless	Evaluated datasets, aggregated (Kedzierska et al. 2025)	Geneformer cell embedding on evaluated datasets (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Rank (rank of the median Average batch score across datasets among the eight representations (1 = best); full embedding dimensions)	8 unitless	Evaluated datasets, aggregated (Kedzierska et al. 2025)	Geneformer cell embedding on evaluated datasets (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Rank (rank of the median Average BIO score across datasets among the eight representations (1 = best); full embedding dimensions)	7 unitless	Evaluated datasets, aggregated (Kedzierska et al. 2025)	Geneformer cell embedding on evaluated datasets (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Rank (rank of the median Batch integration score across datasets among the eight representations (1 = best); full embedding dimensions)	7 unitless	Evaluated datasets, aggregated (Kedzierska et al. 2025)	Geneformer cell embedding on evaluated datasets (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Rank (rank of the median NMI score across datasets among the eight representations (1 = best); full embedding dimensions)	8 unitless	Evaluated datasets, aggregated (Kedzierska et al. 2025)	Geneformer cell embedding on evaluated datasets (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Rank (rank of the median PCR score across datasets among the eight representations (1 = best); full embedding dimensions)	8 unitless	Evaluated datasets, aggregated (Kedzierska et al. 2025)	Geneformer cell embedding on evaluated datasets (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Adjusted rand index (Louvain clusters at best-NMI resolution versus cell-type labels; whole dataset; full embedding dimensions)	0.19 unitless	Immune (330k) (Kedzierska et al. 2025)	Geneformer cell embedding on Immune (330k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Average batch score (average batch score (mean of batch integration score and PCR batch score); whole dataset; full embedding dimensions)	0.38 unitless	Immune (330k) (Kedzierska et al. 2025)	Geneformer cell embedding on Immune (330k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Avg bio (AvgBIO (mean of cell-type ASW, NMI and ARI; NMI and ARI on Louvain clusters at the best-NMI resolution); whole dataset; full embedding dimensions)	0.35 unitless	Immune (330k) (Kedzierska et al. 2025)	Geneformer cell embedding on Immune (330k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Normalized mutual information (Louvain clusters at best-NMI resolution versus cell-type labels; whole dataset; full embedding dimensions)	0.42 unitless	Immune (330k) (Kedzierska et al. 2025)	Geneformer cell embedding on Immune (330k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Rank (rank among the eight representations by mean biological-conservation and batch-mixing scores on this dataset (1 = best); full embedding dimensions)	8 unitless	Immune (330k) (Kedzierska et al. 2025)	Geneformer cell embedding on Immune (330k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Scib asw batch (batch integration score (batch ASW per cell type); whole dataset; full embedding dimensions)	0.76 unitless	Immune (330k) (Kedzierska et al. 2025)	Geneformer cell embedding on Immune (330k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Scib asw label (cell-type ASW; whole dataset; full embedding dimensions)	0.45 unitless	Immune (330k) (Kedzierska et al. 2025)	Geneformer cell embedding on Immune (330k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Scib pcr batch (principal component regression on batch; whole dataset; full embedding dimensions)	0 unitless	Immune (330k) (Kedzierska et al. 2025)	Geneformer cell embedding on Immune (330k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Adjusted rand index (Louvain clusters at best-NMI resolution versus cell-type labels; whole dataset; full embedding dimensions)	0.05 unitless	Pancreas (8k) (Kedzierska et al. 2025)	Geneformer cell embedding on Pancreas (8k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Average batch score (average batch score (mean of batch integration score and PCR batch score); whole dataset; full embedding dimensions)	0.36 unitless	Pancreas (8k) (Kedzierska et al. 2025)	Geneformer cell embedding on Pancreas (8k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Avg bio (AvgBIO (mean of cell-type ASW, NMI and ARI; NMI and ARI on Louvain clusters at the best-NMI resolution); whole dataset; full embedding dimensions)	0.22 unitless	Pancreas (8k) (Kedzierska et al. 2025)	Geneformer cell embedding on Pancreas (8k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Normalized mutual information (Louvain clusters at best-NMI resolution versus cell-type labels; whole dataset; full embedding dimensions)	0.16 unitless	Pancreas (8k) (Kedzierska et al. 2025)	Geneformer cell embedding on Pancreas (8k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Scib asw batch (batch integration score (batch ASW per cell type); whole dataset; full embedding dimensions)	0.73 unitless	Pancreas (8k) (Kedzierska et al. 2025)	Geneformer cell embedding on Pancreas (8k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Scib asw label (cell-type ASW; whole dataset; full embedding dimensions)	0.46 unitless	Pancreas (8k) (Kedzierska et al. 2025)	Geneformer cell embedding on Pancreas (8k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Scib pcr batch (principal component regression on batch; whole dataset; full embedding dimensions)	0 unitless	Pancreas (8k) (Kedzierska et al. 2025)	Geneformer cell embedding on Pancreas (8k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Adjusted rand index (Louvain clusters at best-NMI resolution versus cell-type labels; whole dataset; full embedding dimensions)	0.09 unitless	Pancreas (16k) (Kedzierska et al. 2025)	Geneformer cell embedding on Pancreas (16k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Average batch score (average batch score (mean of batch integration score and PCR batch score); whole dataset; full embedding dimensions)	0.3 unitless	Pancreas (16k) (Kedzierska et al. 2025)	Geneformer cell embedding on Pancreas (16k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Avg bio (AvgBIO (mean of cell-type ASW, NMI and ARI; NMI and ARI on Louvain clusters at the best-NMI resolution); whole dataset; full embedding dimensions)	0.24 unitless	Pancreas (16k) (Kedzierska et al. 2025)	Geneformer cell embedding on Pancreas (16k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Normalized mutual information (Louvain clusters at best-NMI resolution versus cell-type labels; whole dataset; full embedding dimensions)	0.26 unitless	Pancreas (16k) (Kedzierska et al. 2025)	Geneformer cell embedding on Pancreas (16k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Rank (rank among the eight representations by mean biological-conservation and batch-mixing scores on this dataset (1 = best); full embedding dimensions)	8 unitless	Pancreas (16k) (Kedzierska et al. 2025)	Geneformer cell embedding on Pancreas (16k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Scib asw batch (batch integration score (batch ASW per cell type); whole dataset; full embedding dimensions)	0.6 unitless	Pancreas (16k) (Kedzierska et al. 2025)	Geneformer cell embedding on Pancreas (16k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Scib asw label (cell-type ASW; whole dataset; full embedding dimensions)	0.37 unitless	Pancreas (16k) (Kedzierska et al. 2025)	Geneformer cell embedding on Pancreas (16k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Scib pcr batch (principal component regression on batch; whole dataset; full embedding dimensions)	0 unitless	Pancreas (16k) (Kedzierska et al. 2025)	Geneformer cell embedding on Pancreas (16k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Adjusted rand index (Louvain clusters at best-NMI resolution versus cell-type labels; whole dataset; full embedding dimensions)	0.6 unitless	PBMC (12k) (Kedzierska et al. 2025)	Geneformer cell embedding on PBMC (12k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Average batch score (average batch score (mean of batch integration score and PCR batch score); whole dataset; full embedding dimensions)	0.48 unitless	PBMC (12k) (Kedzierska et al. 2025)	Geneformer cell embedding on PBMC (12k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Avg bio (AvgBIO (mean of cell-type ASW, NMI and ARI; NMI and ARI on Louvain clusters at the best-NMI resolution); whole dataset; full embedding dimensions)	0.59 unitless	PBMC (12k) (Kedzierska et al. 2025)	Geneformer cell embedding on PBMC (12k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Normalized mutual information (Louvain clusters at best-NMI resolution versus cell-type labels; whole dataset; full embedding dimensions)	0.65 unitless	PBMC (12k) (Kedzierska et al. 2025)	Geneformer cell embedding on PBMC (12k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Rank (rank among the eight representations by mean biological-conservation and batch-mixing scores on this dataset (1 = best); full embedding dimensions)	8 unitless	PBMC (12k) (Kedzierska et al. 2025)	Geneformer cell embedding on PBMC (12k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Scib asw batch (batch integration score (batch ASW per cell type); whole dataset; full embedding dimensions)	0.97 unitless	PBMC (12k) (Kedzierska et al. 2025)	Geneformer cell embedding on PBMC (12k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Scib asw label (cell-type ASW; whole dataset; full embedding dimensions)	0.51 unitless	PBMC (12k) (Kedzierska et al. 2025)	Geneformer cell embedding on PBMC (12k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Scib pcr batch (principal component regression on batch; whole dataset; full embedding dimensions)	0 unitless	PBMC (12k) (Kedzierska et al. 2025)	Geneformer cell embedding on PBMC (12k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Adjusted rand index (Louvain clusters at best-NMI resolution versus cell-type labels; whole dataset; full embedding dimensions)	0.4 unitless	PBMC (95k) (Kedzierska et al. 2025)	Geneformer cell embedding on PBMC (95k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Avg bio (AvgBIO (mean of cell-type ASW, NMI and ARI; NMI and ARI on Louvain clusters at the best-NMI resolution); whole dataset; full embedding dimensions)	0.52 unitless	PBMC (95k) (Kedzierska et al. 2025)	Geneformer cell embedding on PBMC (95k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Normalized mutual information (Louvain clusters at best-NMI resolution versus cell-type labels; whole dataset; full embedding dimensions)	0.65 unitless	PBMC (95k) (Kedzierska et al. 2025)	Geneformer cell embedding on PBMC (95k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Rank (rank among the eight representations by mean biological-conservation and batch-mixing scores on this dataset (1 = best); full embedding dimensions)	6 unitless	PBMC (95k) (Kedzierska et al. 2025)	Geneformer cell embedding on PBMC (95k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Scib asw label (cell-type ASW; whole dataset; full embedding dimensions)	0.51 unitless	PBMC (95k) (Kedzierska et al. 2025)	Geneformer cell embedding on PBMC (95k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Adjusted rand index (Louvain clusters at best-NMI resolution versus cell-type labels; whole dataset; full embedding dimensions)	0.14 unitless	Tabula Sapiens (483k) (Kedzierska et al. 2025)	Geneformer cell embedding on Tabula Sapiens (483k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Average batch score (average batch score (mean of batch integration score and PCR batch score); whole dataset; full embedding dimensions)	0.45 unitless	Tabula Sapiens (483k) (Kedzierska et al. 2025)	Geneformer cell embedding on Tabula Sapiens (483k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Avg bio (AvgBIO (mean of cell-type ASW, NMI and ARI; NMI and ARI on Louvain clusters at the best-NMI resolution); whole dataset; full embedding dimensions)	0.3 unitless	Tabula Sapiens (483k) (Kedzierska et al. 2025)	Geneformer cell embedding on Tabula Sapiens (483k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Normalized mutual information (Louvain clusters at best-NMI resolution versus cell-type labels; whole dataset; full embedding dimensions)	0.35 unitless	Tabula Sapiens (483k) (Kedzierska et al. 2025)	Geneformer cell embedding on Tabula Sapiens (483k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Rank (rank among the eight representations by mean biological-conservation and batch-mixing scores on this dataset (1 = best); full embedding dimensions)	8 unitless	Tabula Sapiens (483k) (Kedzierska et al. 2025)	Geneformer cell embedding on Tabula Sapiens (483k) (Kedzierska et al. 2025)
Geneformer (Kedzierska et al. 2025)	Scib asw batch (batch integration score (batch ASW per cell type); whole dataset; full embedding dimensions)	0.9 unitless	Tabula Sapiens (483k) (Kedzierska et al. 2025)	Geneformer cell embedding on Tabula Sapiens (483k) (Kedzierska et al. 2025)