Staging installation · synthetic data · not a permanent scientific record · 673dc18 · Merge dev taxonomy work into staging
Circuit Bench

Decoder version

Sparse BP+OSD v0.1

Published

General description

Synthetic accuracy-oriented BP+OSD decoder for interface development.

What changed in this version

Synthetic first revision.

Preparation and output capabilities

Circuit-skeleton preparation
Not required (<10 s on first uncached exposure)
Circuit-prior preparation
Required
Soft output
Provides per-shot failure probability q in [0, 1]

Hyperparameters

Synthetic free-text hyperparameters for UI development.

No machine-readable hyperparameter schema was supplied.

Record history 3 events · revisions and publication decisions
  1. Submitted Ada Decoder Submission event synthesised during history migration. Sparse BP+OSD 0.1
    Event details
    Sequence
    1

    Structured details

    {
  "migration_inferred": true,
  "projected_state": "published"
}

    Canonical submitted snapshot

    {
  "data": {
    "migration_inferred": true,
    "record_id": "8daf39bc-41ee-510e-bf16-58eae7018bed",
    "schema_release_id": "f015bc80-d4ee-52d0-97b6-a380dcd2e1aa"
  },
  "schema": {
    "record_type": "decoder",
    "version": "0.1"
  }
}
  2. Approved System (migration_backfill) Historical approval actor was not recorded. Sparse BP+OSD 0.1
    Event details
    Sequence
    2

    Structured details

    {
  "approval_route": "legacy_unknown",
  "migration_inferred": true
}
  3. Published System (migration_backfill) Publication event synthesised from the lifecycle projection. Sparse BP+OSD 0.1
    Event details
    Sequence
    3
    Caused by
    Event 2

    Structured details

    {
  "migration_inferred": true
}

Results for this decoder version

Scripted query (OData 4.01 subset)
Syntax and field reference
Circuit filters, 0 applied
Code tags all of
No tags selected

Code tags

Selected

No tags selected.

Available tags

Circuit Bench

Error Correction Zoo
Experiment tags all of
No tags selected

Experiment tags

Selected

No tags selected.

Available tags

Circuit Bench
Noise model Any

Choose noise models

Selected

Available noise models

Search results load when the picker opens.
Machine filters, 0 applied

The current population contains 8 published results. Its filters, query, plot, and sort describe the same population.

Show plotter Hide plotter 8 /8 points
LER upper 95% @ 5% acceptance against t₁₀₀₀ 8 of 8 results plotted.
Main plot settings axes, scales and visible ranges

Choose x-axis

Choose y-axis
Advanced plot settings uncertainty, grid and marker appearance
Scatter plot of LER upper 95% @ 5% acceptance against t₁₀₀₀; x is linear and y is linear. Points retain the caller's result order. Stored intervals are displayed as bars. 8 of 8 results have plottable axis values. 4 · 107 6 · 107 8 · 107 1 · 108 1.2 · 108 5 · 10-2 0.1 0.15 t₁₀₀₀ ns · linear LER upper 95% @ 5% acceptance probability · linear
Point details and field definitions

Selected point

Select a point to inspect its result.

x-axis
t_1000_ns — t₁₀₀₀ (ns), linear scale, lower is better. Definition
y-axis
score_ler_upper_95_at_5pct_acceptance_v0_1 — LER upper 95% @ 5% acceptance (probability), linear scale, lower is better. Definition
uncertainty
Whisker bars. 0 x-axis and 8 y-axis stored intervals are available in this plotted population. Only complete lower/upper intervals stored under the selected score definition are drawn; Circuit Bench does not infer missing errors.
Plot data
Tabular data for this plot (8 rows)
LER upper 95% @ 5% acceptance against t₁₀₀₀; omitted points are reported above.
Record Result UUID t₁₀₀₀ (ns) LER upper 95% @ 5% acceptance (probability)
Bivariate bicycle 144 memory 578afcec-cd84-5b24-8549-822adc26e181
Planar stability experiment d=5 d7fae99f-9dfc-58b7-a81f-810ffa1a2809
Rotated surface-code memory d=3 b03fbb8a-58ce-5e2e-a64d-16ff0c74454c 0.14 probability
Rotated surface-code memory d=5 341b6853-2f8a-5865-9710-4f5c14908873 0.1 probability
Rotated surface-code memory d=7 d33dedb2-7e3e-5439-acaf-24bcf9748254
Rotated surface-code memory d=9 29839623-53a4-546a-a4f5-f145b9cff782
Surface-code logical CNOT d=3 89b80853-b04c-5669-b11f-7c977e33a201 0.16 probability
Triangular colour-code memory d=5 116eee62-47db-5c68-9698-e70db02ee588 0.12 probability
8 records Sort: Circuit ascending

Table view options

Choose the columns shown in this table. The selection is recorded in the URL.

Visible columns
Circuit 1↑ Noise model Machine type Machine Shots LER upper 95% @ 5% t₁₀₀₀ (ns) Evaluator scores Reproduction Published
Bivariate bicycle 144 memoryRandomised phenomenological noiseGPUdemo-single-gpu160,000 Brier loss upper 95% bound: 1.75 · 10⁻² probability; LER upper 95% bound at 5% acceptance: 9.5 · 10⁻² probabilityIndependent reproduction31 Aug 2026, 1:10 a.m.
Planar stability experiment d=5Fixed phenomenological noiseGPUdemo-single-gpu140,000 Brier loss upper 95% bound: 1.6 · 10⁻² probability; LER upper 95% bound at 5% acceptance: 8 · 10⁻² probabilityIndependent reproduction31 Aug 2026, 1:08 a.m.
Rotated surface-code memory d=3Fixed phenomenological noiseGPUdemo-single-gpu110,000 0.14 probability Brier loss upper 95% bound: 2.2 · 10⁻² probability; LER upper 95% bound at 5% acceptance: 0.14 probabilityIndependent reproduction31 Aug 2026, 1:05 a.m.
Rotated surface-code memory d=5Fixed phenomenological noiseGPUdemo-single-gpu100,000 0.1 probability Brier loss upper 95% bound: 1.8 · 10⁻² probability; LER upper 95% bound at 5% acceptance: 0.1 probabilityIndependent reproduction31 Aug 2026, 1:04 a.m.
Rotated surface-code memory d=7Randomised phenomenological noiseGPUdemo-single-gpu120,000 Brier loss upper 95% bound: 1.4 · 10⁻² probability; LER upper 95% bound at 5% acceptance: 6 · 10⁻² probabilityIndependent reproduction31 Aug 2026, 1:06 a.m.
Rotated surface-code memory d=9Randomised phenomenological noiseGPUdemo-single-gpu130,000 Brier loss upper 95% bound: 1.175 · 10⁻² probability; LER upper 95% bound at 5% acceptance: 3.75 · 10⁻² probabilityIndependent reproduction31 Aug 2026, 1:07 a.m.
Surface-code logical CNOT d=3Fixed phenomenological noiseGPUdemo-single-gpu170,000 0.16 probability Brier loss upper 95% bound: 2.4 · 10⁻² probability; LER upper 95% bound at 5% acceptance: 0.16 probabilityIndependent reproduction31 Aug 2026, 1:11 a.m.
Triangular colour-code memory d=5Fixed phenomenological noiseGPUdemo-single-gpu150,000 0.12 probability Brier loss upper 95% bound: 2 · 10⁻² probability; LER upper 95% bound at 5% acceptance: 0.12 probabilityIndependent reproduction31 Aug 2026, 1:09 a.m.