EvoFlows, per-family retrained
An EvoFlows editor retrained separately on each target family before generating variants of that family's templates.
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.
Model Capabilities & Hardware Requirements
Deployment facts recorded for EvoFlows, per-family retrained. Fields the record does not carry are marked not recorded.
Evaluated Checkpoint: As printed in the lower block of Table 1
Evaluations and results
12 evaluations · 12 results. Different protocols are not a single leaderboard.
Dataset: The six evaluation datasets behind EvoFlows' published figures
Sorted by Kullback leibler divergence (pooled amino-acid composition, under conditions the replication could not recover) (lower 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 | Kullback leibler divergence (pooled amino-acid composition, under conditions the replication could not recover) | Maximum mean discrepancy (k-mer spectrum MMD, under conditions and a reference set size the replication could not recover) | Covariance agreement (percent of the natural homolog ceiling) | Jensen shannon divergence (positional formulation chosen by the EditJumps authors, expressed as a multiple of the floor) | Mutual information product agreement (percent of the natural homolog ceiling) | Edit count (mean edits per sequence as the EvoFlows paper prints it, under conditions the replication could not recover) | Levenshtein distance (printed under the column heading 'diversity'; mean pairwise Levenshtein distance, under conditions the replication could not recover) | Levenshtein distance (pooled reduction, expressed as a ratio to natural homologs) |
|---|---|---|---|---|---|---|---|---|---|
| EvoFlows, per-family retrainedEvaluation | EvoFlows published values, two families: KL x10^3 | 21.5 (best in column) | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported |
| EvoFlows, per-family retrainedEvaluation | EvoFlows published values, two families: KL x10^3 | 23.2 | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported |
| EvoFlows, per-family retrainedEvaluation | Ceiling-normalised comparison, EvoFlows published values, six families: Covariance (% of ceiling) | Not reported | Not reported | 98.4% | Not reported | Not reported | Not reported | Not reported | Not reported |
| EvoFlows, per-family retrainedEvaluation | Ceiling-normalised comparison, EvoFlows published values, six families: Positional JS† (× floor) | Not reported | Not reported | Not reported | 1.8 | Not reported | Not reported | Not reported | Not reported |
| EvoFlows, per-family retrainedEvaluation | Ceiling-normalised comparison, EvoFlows published values, six families: MIP (% of ceiling) | Not reported | Not reported | Not reported | Not reported | 95.5% | Not reported | Not reported | Not reported |
| EvoFlows, per-family retrainedEvaluation | Ceiling-normalised comparison, EvoFlows published values, six families: Pairwise Distance‡ | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported | 1 |
| EvoFlows, per-family retrainedEvaluation | EvoFlows published values, two families: edits/seq | Not reported | Not reported | Not reported | Not reported | Not reported | 35.3 | Not reported | Not reported |
| EvoFlows, per-family retrainedEvaluation | EvoFlows published values, two families: edits/seq | Not reported | Not reported | Not reported | Not reported | Not reported | 14.6 | Not reported | Not reported |
| EvoFlows, per-family retrainedEvaluation | EvoFlows published values, two families: diversity | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported | 90.9 | Not reported |
| EvoFlows, per-family retrainedEvaluation | EvoFlows published values, two families: diversity | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported | 35.6 | Not reported |
| EvoFlows, per-family retrainedEvaluation | EvoFlows published values, two families: MMD | Not reported | 1.57 | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported |
| EvoFlows, per-family retrainedEvaluation | EvoFlows published values, two families: MMD | Not reported | 1.03 (best in column) | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported |
All 12 result rows with coverage, uncertainty and sources
Filter evaluations
Applied filters: All linked evaluations
| Tested configuration | Protocol and dataset | Finding | Evidence and details |
|---|---|---|---|
| Configuration: EvoFlows, per-family retrained | Protocol: Ceiling-normalised comparison, EvoFlows published values, six families: Covariance (% of ceiling) Dataset: The six evaluation datasets behind EvoFlows' published figures | 98.4% Covariance agreement (percent of the natural homolog ceiling) percent · higher Uncertainty: Not reported by the source: The table prints a single value with no interval. Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · Source checkedMethods, coverage and sourceEvoFlows (Published), Covariance (% of ceiling), ceiling-normalised Metrics normalised against the attainable ceiling computed on held-out natural homologs. Aggregation: Not reported When Edit Flows are Edit Jumps: replicating Edit Flows and EvoFlows (arXiv v1) · Table 4, row "Covariance (% of ceiling)", column "EvoFlows (Published)" |
| Configuration: EvoFlows, per-family retrained | Protocol: Ceiling-normalised comparison, EvoFlows published values, six families: Positional JS† (× floor) Dataset: The six evaluation datasets behind EvoFlows' published figures | 1.8 Jensen shannon divergence (positional formulation chosen by the EditJumps authors, expressed as a multiple of the floor) unitless · lower Uncertainty: Not reported by the source: The table prints a single value with no interval. Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · Source checkedMethods, coverage and sourceEvoFlows (Published), Positional JS† (× floor), ceiling-normalised Metrics normalised against the attainable ceiling computed on held-out natural homologs. Aggregation: Not reported When Edit Flows are Edit Jumps: replicating Edit Flows and EvoFlows (arXiv v1) · Table 4, row "Positional JS† (× floor)", column "EvoFlows (Published)" |
| Configuration: EvoFlows, per-family retrained | Protocol: Ceiling-normalised comparison, EvoFlows published values, six families: MIP (% of ceiling) Dataset: The six evaluation datasets behind EvoFlows' published figures | 95.5% Mutual information product agreement (percent of the natural homolog ceiling) percent · higher Uncertainty: Not reported by the source: The table prints a single value with no interval. Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · Source checkedMethods, coverage and sourceEvoFlows (Published), MIP (% of ceiling), ceiling-normalised Metrics normalised against the attainable ceiling computed on held-out natural homologs. Aggregation: Not reported When Edit Flows are Edit Jumps: replicating Edit Flows and EvoFlows (arXiv v1) · Table 4, row "MIP (% of ceiling)", column "EvoFlows (Published)" |
| Configuration: EvoFlows, per-family retrained | Protocol: Ceiling-normalised comparison, EvoFlows published values, six families: Pairwise Distance‡ Dataset: The six evaluation datasets behind EvoFlows' published figures | 1 Levenshtein distance (pooled reduction, expressed as a ratio to natural homologs) unitless · unknown Uncertainty: Not reported by the source: The table prints a single value with no interval. Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · Source checkedMethods, coverage and sourceEvoFlows (Published), Pairwise Distance‡, ceiling-normalised Metrics normalised against the attainable ceiling computed on held-out natural homologs. Aggregation: Not reported When Edit Flows are Edit Jumps: replicating Edit Flows and EvoFlows (arXiv v1) · Table 4, row "Pairwise Distance‡", column "EvoFlows (Published)" |
| Configuration: EvoFlows, per-family retrained | Protocol: EvoFlows published values, two families: edits/seq Dataset: The six evaluation datasets behind EvoFlows' published figures | 35.3 Edit count (mean edits per sequence as the EvoFlows paper prints it, under conditions the replication could not recover) count · unknown Uncertainty: Not reported by the source: The table prints a single value with no interval. Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · Source checkedMethods, coverage and sourceEvoFlows on HER2-VH, edits/seq, EvoFlows published values The conditions under which the EvoFlows paper produced these values, which the replication states it could not recover. Aggregation: Not reported When Edit Flows are Edit Jumps: replicating Edit Flows and EvoFlows (arXiv v1) · Table 1, lower block, row "EvoFlows", column "edits/seq", sub-column "HER2" |
| Configuration: EvoFlows, per-family retrained | Protocol: EvoFlows published values, two families: edits/seq Dataset: The six evaluation datasets behind EvoFlows' published figures | 14.6 Edit count (mean edits per sequence as the EvoFlows paper prints it, under conditions the replication could not recover) count · unknown Uncertainty: Not reported by the source: The table prints a single value with no interval. Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · Source checkedMethods, coverage and sourceEvoFlows on Ty1, edits/seq, EvoFlows published values The conditions under which the EvoFlows paper produced these values, which the replication states it could not recover. Aggregation: Not reported When Edit Flows are Edit Jumps: replicating Edit Flows and EvoFlows (arXiv v1) · Table 1, lower block, row "EvoFlows", column "edits/seq", sub-column "Ty1" |
| Configuration: EvoFlows, per-family retrained | Protocol: EvoFlows published values, two families: KL x10^3 Dataset: The six evaluation datasets behind EvoFlows' published figures | 23.2 Kullback leibler divergence (pooled amino-acid composition, under conditions the replication could not recover) unitless · lower Uncertainty: Not reported by the source: The table prints a single value with no interval. Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · Source checkedMethods, coverage and sourceEvoFlows on HER2-VH, KL x10^3, EvoFlows published values The conditions under which the EvoFlows paper produced these values, which the replication states it could not recover. Aggregation: Not reported When Edit Flows are Edit Jumps: replicating Edit Flows and EvoFlows (arXiv v1) · Table 1, lower block, row "EvoFlows", column "KL ×103", sub-column "HER2" |
| Configuration: EvoFlows, per-family retrained | Protocol: EvoFlows published values, two families: KL x10^3 Dataset: The six evaluation datasets behind EvoFlows' published figures | 21.5 Kullback leibler divergence (pooled amino-acid composition, under conditions the replication could not recover) unitless · lower Uncertainty: Not reported by the source: The table prints a single value with no interval. Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · Source checkedMethods, coverage and sourceEvoFlows on Ty1, KL x10^3, EvoFlows published values The conditions under which the EvoFlows paper produced these values, which the replication states it could not recover. Aggregation: Not reported When Edit Flows are Edit Jumps: replicating Edit Flows and EvoFlows (arXiv v1) · Table 1, lower block, row "EvoFlows", column "KL ×103", sub-column "Ty1" |
| Configuration: EvoFlows, per-family retrained | Protocol: EvoFlows published values, two families: diversity Dataset: The six evaluation datasets behind EvoFlows' published figures | 90.9 Levenshtein distance (printed under the column heading 'diversity'; mean pairwise Levenshtein distance, under conditions the replication could not recover) count · unknown Uncertainty: Not reported by the source: The table prints a single value with no interval. Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · Source checkedMethods, coverage and sourceEvoFlows on HER2-VH, diversity, EvoFlows published values The conditions under which the EvoFlows paper produced these values, which the replication states it could not recover. Aggregation: Not reported When Edit Flows are Edit Jumps: replicating Edit Flows and EvoFlows (arXiv v1) · Table 1, lower block, row "EvoFlows", column "diversity", sub-column "HER2" |
| Configuration: EvoFlows, per-family retrained | Protocol: EvoFlows published values, two families: diversity Dataset: The six evaluation datasets behind EvoFlows' published figures | 35.6 Levenshtein distance (printed under the column heading 'diversity'; mean pairwise Levenshtein distance, under conditions the replication could not recover) count · unknown Uncertainty: Not reported by the source: The table prints a single value with no interval. Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · Source checkedMethods, coverage and sourceEvoFlows on Ty1, diversity, EvoFlows published values The conditions under which the EvoFlows paper produced these values, which the replication states it could not recover. Aggregation: Not reported When Edit Flows are Edit Jumps: replicating Edit Flows and EvoFlows (arXiv v1) · Table 1, lower block, row "EvoFlows", column "diversity", sub-column "Ty1" |
| Configuration: EvoFlows, per-family retrained | Protocol: EvoFlows published values, two families: MMD Dataset: The six evaluation datasets behind EvoFlows' published figures | 1.57 Maximum mean discrepancy (k-mer spectrum MMD, under conditions and a reference set size the replication could not recover) unitless · lower Uncertainty: Not reported by the source: The table prints a single value with no interval. Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · Source checkedMethods, coverage and sourceEvoFlows on HER2-VH, MMD, EvoFlows published values The conditions under which the EvoFlows paper produced these values, which the replication states it could not recover. Aggregation: Not reported When Edit Flows are Edit Jumps: replicating Edit Flows and EvoFlows (arXiv v1) · Table 1, lower block, row "EvoFlows", column "MMD", sub-column "HER2" |
| Configuration: EvoFlows, per-family retrained | Protocol: EvoFlows published values, two families: MMD Dataset: The six evaluation datasets behind EvoFlows' published figures | 1.03 Maximum mean discrepancy (k-mer spectrum MMD, under conditions and a reference set size the replication could not recover) unitless · lower Uncertainty: Not reported by the source: The table prints a single value with no interval. Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · Source checkedMethods, coverage and sourceEvoFlows on Ty1, MMD, EvoFlows published values The conditions under which the EvoFlows paper produced these values, which the replication states it could not recover. Aggregation: Not reported When Edit Flows are Edit Jumps: replicating Edit Flows and EvoFlows (arXiv v1) · Table 1, lower block, row "EvoFlows", column "MMD", sub-column "Ty1" |
Source checking is not independent reproduction. Release 2026-11-21-c8bea25e5abb.
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
| Property and statement | Original source and location | Review and provenance |
|---|
No evidence rows match these filters. Choose another scope or clear the search.
Sources and history
Release 2026-11-21-c8bea25e5abb · Record review: source checked
1 source record and release history
- When Edit Flows are Edit Jumps: replicating Edit Flows and EvoFlows (arXiv v1) · Original source · arXiv:2609.18745v1, 16 September 2026
Technical metadata and record identifiers
Technical metadata for this record is tracked under stable ID editjumps-20261115-config-published-evoflows.
Source locator: Table 1, lower block, row 'EvoFlows'
Related records
- subject: There is no retrained EvoFlows to compare against
- subject: The published comparison values were recovered from a figure, not a table
- subject: The two sets of numbers are not comparable
- system: EvoFlows (Published), Covariance (% of ceiling), ceiling-normalised
- system: EvoFlows (Published), Positional JS† (× floor), ceiling-normalised
- system: EvoFlows (Published), MIP (% of ceiling), ceiling-normalised
- system: EvoFlows (Published), Pairwise Distance‡, ceiling-normalised
- system: EvoFlows on HER2-VH, edits/seq, EvoFlows published values
- system: EvoFlows on Ty1, edits/seq, EvoFlows published values
- system: EvoFlows on HER2-VH, KL x10^3, EvoFlows published values
- system: EvoFlows on Ty1, KL x10^3, EvoFlows published values
- system: EvoFlows on HER2-VH, diversity, EvoFlows published values
- system: EvoFlows on Ty1, diversity, EvoFlows published values
- system: EvoFlows on HER2-VH, MMD, EvoFlows published values
- system: EvoFlows on Ty1, MMD, EvoFlows published values
How to cite
EvoFlows, per-family retrained. Rewire biological model benchmark database, release 2026-11-21-c8bea25e5abb. https://benchmarks.rewirebio.io/database/configuration/editjumps-20261115-config-published-evoflows/
Release 2026-11-21-c8bea25e5abb. Rankings and values change between releases, so keep the release id in the citation. Cite the original publication as well for any value you reuse: each result links its source.
The compilation and annotations made by Rewire in this release are offered under CC BY 4.0; values and text taken from cited sources stay under the terms of those sources and are not covered by this licence. Data licence