rewire.it
Task

GENEB LINEAR-PROBE: Average macro-MCC across the 13 representative tasks, linear probe

Average macro-MCC across the 13 representative tasks, linear probe. Scored with Macro-MCC on GENEB representative task subset. Frozen embeddings scored with a probe over the thirteen representative tasks, one from each functional category (Table 7).

11 evaluations · 11 metric rows

Overview

Average macro-MCC across the 13 representative tasks, linear probe. Scored with Macro-MCC on GENEB representative task subset. Frozen embeddings scored with a probe over the thirteen representative tasks, one from each functional category (Table 7).

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

Evaluation design

Benchmarks bring together tasks and protocols. A task describes the biological question; a protocol defines a particular test.

Benchmarks

These source-backed links do not make different protocols or scores interchangeable.

Recorded 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.

Published comparisons

Explore the results reported under one evaluation protocol. Each figure keeps its source, dataset and metric together; it is not a ranking across studies. The pooled view gathers every source table that reports the same metric and names what it does not hold constant.

GENEB LINEAR-PROBE: Average macro-MCC across the 13 representative tasks, linear probe

macro_mcc (correlation) · Higher values are better for this metric.

Every method GENEB reports on Average macro-MCC across the 13 representative tasks, linear probe, scored with Macro-MCC on GENEB representative task subset.

Evaluation protocol · GENEB representative task subset (GENEB split)

  1. GENERator-Eukaryote-3B · Configuration · Author-reported evaluation0.605
  2. LucaOne · Configuration · Author-reported evaluation0.573
  3. GenomeOcean-4B · Configuration · Author-reported evaluation0.552
  4. Omni-DNA-1B · Configuration · Author-reported evaluation0.550
  5. GenomeOcean-500M · Configuration · Author-reported evaluation0.536
  6. GENA-LM-Large-T2T · Configuration · Author-reported evaluation0.530
  7. MutBERT · Configuration · Author-reported evaluation0.516
  8. NT-v2-50M-MS · Configuration · Author-reported evaluation0.511
  9. GROVER · Configuration · Author-reported evaluation0.466
  10. HyenaDNA-Large-1M · Configuration · Author-reported evaluation0.427

Source order is preserved. Plotted marks show point estimates; uncertainty, where reported, is retained in the printed values and table. Differences do not establish statistical significance.

GENEB: Why Genomic Models Are Hard to Compare · Table 8, column(Linear MCC)
Values, uncertainty and evidence
macro_mcc: original source values
Tested entityPrinted valueUncertaintyEvidence
GENERator-Eukaryote-3B · Configuration0.605 correlationNot reportedAuthor-reported evaluation · source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(GENERator-Eukaryote-3B), column(Linear MCC)
GENERator-Eukaryote-1.2B · Configuration0.579 correlationNot reportedAuthor-reported evaluation · source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(GENERator-Eukaryote-1.2B), column(Linear MCC)
LucaOne · Configuration0.573 correlationNot reportedAuthor-reported evaluation · source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(LucaOne), column(Linear MCC)
GenomeOcean-4B · Configuration0.552 correlationNot reportedAuthor-reported evaluation · source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(GenomeOcean-4B), column(Linear MCC)
Omni-DNA-1B · Configuration0.550 correlationNot reportedAuthor-reported evaluation · source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(Omni-DNA-1B), column(Linear MCC)
GenomeOcean-500M · Configuration0.536 correlationNot reportedAuthor-reported evaluation · source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(GenomeOcean-500M), column(Linear MCC)
GENA-LM-Large-T2T · Configuration0.530 correlationNot reportedAuthor-reported evaluation · source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(GENA-LM-Large-T2T), column(Linear MCC)
MutBERT · Configuration0.516 correlationNot reportedAuthor-reported evaluation · source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(MutBERT), column(Linear MCC)
NT-v2-50M-MS · Configuration0.511 correlationNot reportedAuthor-reported evaluation · source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(NT-v2-50M-MS), column(Linear MCC)
GROVER · Configuration0.466 correlationNot reportedAuthor-reported evaluation · source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(GROVER), column(Linear MCC)
HyenaDNA-Large-1M · Configuration0.427 correlationNot reportedAuthor-reported evaluation · source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(HyenaDNA-Large-1M), column(Linear MCC)
Scope and limitations
  • Author-reported numbers, source checked but not independently reproduced.
  • These are averages over the thirteen representative tasks, not a score on any single task.
  • The paper's own point is that these rankings shift with the probe and the protocol, so a position here is not a general ranking.

Source transcription and grouping reviewed by automated source review on 2026-09-18. These experiments were not independently reproduced by rewire.

Tested entities and results

Release 2026-09-17-134cd1815de8 · 11 evaluations · 11 metric rows. Different protocols are not a single leaderboard. Where several source tables report the same metric, the published comparisons above offer a pooled view that names what it does not hold constant.

Results grouped by the exact reported evaluation
Metric and findingCoverage and uncertaintyEvidence
GENA-LM-Large-T2T on GENEB LINEAR-PROBE: Average macro-MCC across the 13 representative tasks, linear probe

Frozen embeddings scored with a probe over the thirteen representative tasks, one from each functional category (Table 7).

Author-reported evaluation · Evaluation metadata: source checked

0.530 macro_mcc

Unit: correlation · Direction: higher

Uncertainty: Not reported

Scored: Not reported · Eligible: Not reported

source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(GENA-LM-Large-T2T), column(Linear MCC)

Source checking is not independent reproduction.

GENERator-Eukaryote-1.2B on GENEB LINEAR-PROBE: Average macro-MCC across the 13 representative tasks, linear probe

Frozen embeddings scored with a probe over the thirteen representative tasks, one from each functional category (Table 7).

Author-reported evaluation · Evaluation metadata: source checked

0.579 macro_mcc

Unit: correlation · Direction: higher

Uncertainty: Not reported

Scored: Not reported · Eligible: Not reported

source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(GENERator-Eukaryote-1.2B), column(Linear MCC)

Source checking is not independent reproduction.

GENERator-Eukaryote-3B on GENEB LINEAR-PROBE: Average macro-MCC across the 13 representative tasks, linear probe

Frozen embeddings scored with a probe over the thirteen representative tasks, one from each functional category (Table 7).

Author-reported evaluation · Evaluation metadata: source checked

0.605 macro_mcc

Unit: correlation · Direction: higher

Uncertainty: Not reported

Scored: Not reported · Eligible: Not reported

source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(GENERator-Eukaryote-3B), column(Linear MCC)

Source checking is not independent reproduction.

GenomeOcean-4B on GENEB LINEAR-PROBE: Average macro-MCC across the 13 representative tasks, linear probe

Frozen embeddings scored with a probe over the thirteen representative tasks, one from each functional category (Table 7).

Author-reported evaluation · Evaluation metadata: source checked

0.552 macro_mcc

Unit: correlation · Direction: higher

Uncertainty: Not reported

Scored: Not reported · Eligible: Not reported

source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(GenomeOcean-4B), column(Linear MCC)

Source checking is not independent reproduction.

GenomeOcean-500M on GENEB LINEAR-PROBE: Average macro-MCC across the 13 representative tasks, linear probe

Frozen embeddings scored with a probe over the thirteen representative tasks, one from each functional category (Table 7).

Author-reported evaluation · Evaluation metadata: source checked

0.536 macro_mcc

Unit: correlation · Direction: higher

Uncertainty: Not reported

Scored: Not reported · Eligible: Not reported

source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(GenomeOcean-500M), column(Linear MCC)

Source checking is not independent reproduction.

GROVER on GENEB LINEAR-PROBE: Average macro-MCC across the 13 representative tasks, linear probe

Frozen embeddings scored with a probe over the thirteen representative tasks, one from each functional category (Table 7).

Author-reported evaluation · Evaluation metadata: source checked

0.466 macro_mcc

Unit: correlation · Direction: higher

Uncertainty: Not reported

Scored: Not reported · Eligible: Not reported

source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(GROVER), column(Linear MCC)

Source checking is not independent reproduction.

HyenaDNA-Large-1M on GENEB LINEAR-PROBE: Average macro-MCC across the 13 representative tasks, linear probe

Frozen embeddings scored with a probe over the thirteen representative tasks, one from each functional category (Table 7).

Author-reported evaluation · Evaluation metadata: source checked

0.427 macro_mcc

Unit: correlation · Direction: higher

Uncertainty: Not reported

Scored: Not reported · Eligible: Not reported

source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(HyenaDNA-Large-1M), column(Linear MCC)

Source checking is not independent reproduction.

LucaOne on GENEB LINEAR-PROBE: Average macro-MCC across the 13 representative tasks, linear probe

Frozen embeddings scored with a probe over the thirteen representative tasks, one from each functional category (Table 7).

Author-reported evaluation · Evaluation metadata: source checked

0.573 macro_mcc

Unit: correlation · Direction: higher

Uncertainty: Not reported

Scored: Not reported · Eligible: Not reported

source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(LucaOne), column(Linear MCC)

Source checking is not independent reproduction.

MutBERT on GENEB LINEAR-PROBE: Average macro-MCC across the 13 representative tasks, linear probe

Frozen embeddings scored with a probe over the thirteen representative tasks, one from each functional category (Table 7).

Author-reported evaluation · Evaluation metadata: source checked

0.516 macro_mcc

Unit: correlation · Direction: higher

Uncertainty: Not reported

Scored: Not reported · Eligible: Not reported

source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(MutBERT), column(Linear MCC)

Source checking is not independent reproduction.

NT-v2-50M-MS on GENEB LINEAR-PROBE: Average macro-MCC across the 13 representative tasks, linear probe

Frozen embeddings scored with a probe over the thirteen representative tasks, one from each functional category (Table 7).

Author-reported evaluation · Evaluation metadata: source checked

0.511 macro_mcc

Unit: correlation · Direction: higher

Uncertainty: Not reported

Scored: Not reported · Eligible: Not reported

source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(NT-v2-50M-MS), column(Linear MCC)

Source checking is not independent reproduction.

Omni-DNA-1B on GENEB LINEAR-PROBE: Average macro-MCC across the 13 representative tasks, linear probe

Frozen embeddings scored with a probe over the thirteen representative tasks, one from each functional category (Table 7).

Author-reported evaluation · Evaluation metadata: source checked

0.550 macro_mcc

Unit: correlation · Direction: higher

Uncertainty: Not reported

Scored: Not reported · Eligible: Not reported

source checkedGENEB: Why Genomic Models Are Hard to Compare · Table 8, row(Omni-DNA-1B), column(Linear MCC)

Source checking is not 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 2026-09-17-134cd1815de8
Property and statementOriginal source and locationReview and provenance
Relationship: part of

discovery-benchmark-geneb

Individual claims
GENEB: Why Genomic Models Are Hard to Compare

Original source ↗

Table 8, column(Linear MCC)

Version: 2606.04525v1
Retrieved: 2026-09-17T07:56:09.460928+00:00

source checked

automated source review · 2026-09-18

Audit details

Primary-source transcription with no human sign-off and no independent reproduction.

Field: links:part_of:discovery-benchmark-geneb

Claim: geneb-association-linear-probe

Source artifact SHA-256: 47975089c0ca738d5e2d6e6ea91cd4e7b80498c3f175803d8ec39ca77aa41629

Hash scope: Exact retrieved primary paper artifact bytes.

Inspected artifact

Sources and history

Release 2026-09-17-134cd1815de8 · Record review: source checked

1 source records and release historyDownload this release
Technical metadata and extraction receipts

Stable ID: geneb-task-linear-probe

areas
dna-genomes
tasks
Average macro-MCC across the 13 representative tasks, linear probe
metric
Macro-MCC
metric direction
higher
dataset
GENEB representative task subset
protocol
Frozen embeddings scored with a probe over the thirteen representative tasks, one from each functional category (Table 7).
source locator
Table 8, column(Linear MCC)
comparison panels
id: geneb-panel-linear-probe; title: GENEB LINEAR-PROBE: Average macro-MCC across the 13 representative tasks, linear probe; protocol id: geneb-task-linear-probe; dataset id: geneb-dataset-geneb-representative-task-subset; metric: macro_mcc; unit: correlation; direction: higher; result ids: geneb-result-generator-eukaryote-3b-linear-probe-macro-mcc; geneb-result-generator-eukaryote-1-2b-linear-probe-macro-mcc; geneb-result-lucaone-linear-probe-macro-mcc; geneb-result-genomeocean-4b-linear-probe-macro-mcc; geneb-result-omni-dna-1b-linear-probe-macro-mcc; geneb-result-genomeocean-500m-linear-probe-macro-mcc; geneb-result-gena-lm-large-t2t-linear-probe-macro-mcc; geneb-result-mutbert-linear-probe-macro-mcc; geneb-result-nt-v2-50m-ms-linear-probe-macro-mcc; geneb-result-grover-linear-probe-macro-mcc; geneb-result-hyenadna-large-1m-linear-probe-macro-mcc; source ids: evidence-expansion-p2-evidence-discovery-final-geneb-47975089c0ca; source locator: Table 8, column(Linear MCC); context: Every method GENEB reports on Average macro-MCC across the 13 representative tasks, linear probe, scored with Macro-MCC on GENEB representative task subset.; caveats: Author-reported numbers, source checked but not independently reproduced.; These are averages over the thirteen representative tasks, not a score on any single task.; The paper's own point is that these rankings shift with the probe and the protocol, so a position here is not a general ranking.; review: method: automated_source_review; date: 2026-09-18
Related records

Suggest a correction