Datasets
Regulatory-edge evaluation across tissue domains; exact dataset releases remain unextracted.
Task
gene-regulatory signal prediction
Gene-regulatory signal prediction tests whether combined representations generalize beyond repeated gene identities.
Sources
Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passagesOverview
Metrics
AUROC in the reviewed stacked-model comparison.
Allowed inputs
Candidate transcription-factor/target pairs and model representations.
Sources
Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passagesEvaluation procedure diagram
How it worksComputational evaluation flow
Conceptual summary of the cited evaluation; exact task configuration and source version remain part of the protocol.
Sources
Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passageslimited source coverage · Automated source review, 2026-09-16. All specifications and missing details
Results
Results are available, but no reviewed comparison panel is linked in this release.
All evaluations
1 evaluation · 1 result. Different protocols are not a single leaderboard.
Filter evaluations
Applied filters: All linked evaluations
| Tested configuration | Protocol and dataset | Finding | Evidence and details |
|---|---|---|---|
| Pipeline: scGPT + residual geometry | Task: gene-regulatory signal prediction Dataset: immune tissue | 0.677 AUROC fraction · unknown Uncertainty: Not reported Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · Source checkedMethods, coverage and sourcescGPT + residual geometry: gene-regulatory signal prediction Asymmetric extraction, PCA-64 centered cosine geometry added to scGPT baseline Aggregation: Not reported Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Table 4, Immune row, scGPT > +geom AUROC column |
Source checking is not independent reproduction. Release 2026-09-29-06401fd5b220.
Methods and evaluation design
Procedure, tasks and evaluated configurations
How it works
Evaluation methodology
Regulatory-edge evaluation across tissue domains; exact dataset releases remain unextracted. Nested outer-split evaluation is complemented by leave-TF-out, leave-target-out and leave-both-out tests. AUROC in the reviewed stacked-model comparison. Combined representation is compared with the individual model branches. Nested validation separates feature construction from testing; entity-grouped tests address gene-identity carryover. Bootstrap confidence intervals support reported comparisons.
Sources
Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passagesRecorded evaluations
Each evaluation records what was tested and under which conditions.
Run instructions
No runnable recipe has been reviewed for this task. Dataset access, model requirements, licences and compute requirements must be checked against its sources before execution.
A task describes a biological question. Choose a linked protocol to obtain concrete split and scoring instructions.
Strengths, limitations and unresolved questions
Strengths and limitations
Strengths and considerations
- Separate TF, target and joint holdouts diagnose which entity-sharing shortcuts remain possible.
Sources
Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages
Limitations and conditions
- Random edge splits and entity-held-out splits answer different questions. Exact bootstrap unit and reference-edge uncertainty remain unextracted.
Sources
Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages
Profile review details
Task-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced.
Stable record: reported-task-99afd88cb12895Specifications
Inputs, training, access and other details
Explanatory profile: limited source coverage · Automated source review, 2026-09-16. Review applies to the cited claims; unresolved fields are listed below. Numerical results retain their own review status.
| Property | Description and evidence |
|---|---|
| Datasets | Regulatory-edge evaluation across tissue domains; exact dataset releases remain unextracted.SourcesResidual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages |
| Splits | Nested outer-split evaluation is complemented by leave-TF-out, leave-target-out and leave-both-out tests.SourcesResidual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages |
| Metrics | AUROC in the reviewed stacked-model comparison.SourcesResidual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages |
| Baselines | Combined representation is compared with the individual model branches.SourcesResidual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages |
| Leakage controls | Nested validation separates feature construction from testing; entity-grouped tests address gene-identity carryover.SourcesResidual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages |
| Uncertainty | Bootstrap confidence intervals support reported comparisons.SourcesResidual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages |
| Entity type | Paper-specific computational evaluation protocol.SourcesResidual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages |
| Organisms | Human tissues from the Tabula Sapiens atlas.SourcesResidual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages |
| Assays | Tabula Sapiens single-cell RNA-seq with TRRUST regulatory reference annotations.SourcesResidual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages |
| Allowed inputs | Candidate transcription-factor/target pairs and model representations.SourcesResidual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages |
| Adaptation | Supervised prediction with nested splits and TF/target holdout stress tests.SourcesResidual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages |
Evidence
Source checking verifies the cited claim or transcription. It does not establish independent reproduction.
Papers and result coverage
Last literature check: 2026-09-17. Primary-paper discovery and source inspection. Source-checked results are not independently reproduced experiments.
| Paper or primary resource | Version | Reference |
|---|---|---|
| Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues | version of record | Read source DOI: 10.1186/s12859-026-06538-5 |
Historical gaps recorded on 2026-09-17
The catalogue now holds 1 result rows for this benchmark. A note below about pending extraction describes the state on 2026-09-17 and may since have been answered by a later batch. The result rows and their sources are the current record.
- Original asymmetric extraction in Table 4 is not a fair matched-input comparison; prefer Table 9 with its single-seed context.
- Strict leave-both-out improvements shrink near zero; do not omit this limitation.
- Source contains rounded deltas and intervals/significance annotations; do not recompute them from rounded absolute numbers.
Search and extraction details
primary comparison table screened
Searches
- "PMC13418759"
Evidence locations
- Tables 7 and 9 and footnotes
- Results: matched per-layer extraction
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.
18 evidence rows matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
| Diagram caption Conceptual summary of the cited evaluation; exact task configuration and source version remain part of the protocol. Individual claims | Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
Diagram steps
| Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Diagram title Computational evaluation flow Individual claims | Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Datasets Regulatory-edge evaluation across tissue domains; exact dataset releases remain unextracted. Individual claims | Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Splits Nested outer-split evaluation is complemented by leave-TF-out, leave-target-out and leave-both-out tests. Individual claims | Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Adaptation Supervised prediction with nested splits and TF/target holdout stress tests. Individual claims | Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Metrics AUROC in the reviewed stacked-model comparison. Individual claims | Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Baselines Combined representation is compared with the individual model branches. Individual claims | Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Leakage controls Nested validation separates feature construction from testing; entity-grouped tests address gene-identity carryover. Individual claims | Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Uncertainty Bootstrap confidence intervals support reported comparisons. Individual claims | Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages Version: version of record | source checked automated source review · 2026-09-16 Audit detailsTask-specific computational methodology and field context checked in the cited primary-source artifact. Source-backed fields, inapplicable evaluator dimensions and unresolved details are distinguished. Numerical results were not reproduced. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
Sources and history
View linked audit checks and correction history
Release 2026-09-29-06401fd5b220 · Record review: needs review
2 source records and release history
- Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Original source · version of record
- Residual-stream geometry of single-cell foundation models carries incremental gene-regulatory signal across tissues · Original source · version of record
Technical metadata and extraction receipts
Stable ID: reported-task-99afd88cb12895
- areas
- cells-tissues
- tasks
- gene-regulatory signal prediction
- entity level
- task
- version
- Not reported
- task
- gene-regulatory signal prediction
- scope note
- Paper-specific evaluation task; protocol completeness requires further extraction.
- benchmark research
- review date: 2026-09-17; status: primary_comparison_table_screened; primary sources: expansion-p3-single-cell-residual-geometry-2026; inspected locators: Tables 7 and 9 and footnotes; Results: matched per-layer extraction; searched queries: "PMC13418759"; gaps: Original asymmetric extraction in Table 4 is not a fair matched-input comparison; prefer Table 9 with its single-seed context.; Strict leave-both-out improvements shrink near zero; do not omit this limitation.; Source contains rounded deltas and intervals/significance annotations; do not recompute them from rounded absolute numbers.; claim scope: Primary-paper discovery and source inspection. Source-checked results are not independently reproduced experiments.
- historical missing metadata
- protocol version: not_reported_in_legacy_extract; split: not_reported_in_legacy_extract
- metadata review scope
- historical_missing_metadata preserves the original discovery state. Current descriptive evidence and missingness are recorded in profile.facts; numerical-result review is separate.
- legacy kinds
- benchmark
- entity classification
- review date: 2026-09-17; rationale: This source-scoped record identifies the biological prediction task and holds its paper context. Preserve the existing task identity; exact split, model adaptation and scoring remain in linked evaluations or separate protocol records.; source ids: single-cell-residual-geometry-2026; source locator: Results: nested cross-validation; Grouped cross-validation; cached text lines 48–56; Methods dataset and scGPT passages; ambiguities: A paper- or suite-specific task may constrain some inputs or metrics; that alone does not make it interchangeable with a complete versioned protocol. No protocol equivalence is inferred.; Some legacy profile Entity type facts use the generic phrase computational evaluation protocol. That boilerplate is not sufficient to establish a single fixed protocol identity or to merge this task with another protocol record.
Related records
- dataset: immune tissue
- benchmark: scGPT + residual geometry: gene-regulatory signal prediction