Quantum Reality Lab Experiments & evidence

Research investigations · 5

A headline is a starting point.
Make it an investigation.

Ask a question. Make a prediction. Compare exact theory with finite samples—and see where the model stops.

Reduced-scale repetition-code model · small-system teaching model

Can parity checks locate a flipped bit?

Explore a small building block related to the HRL team’s silicon-QPU code experiments.

  1. 1Predict
  2. 2Ideal
  3. 3Ideal Shots
  4. 4Hardware
  5. 5Interpret
  6. 6Lab note
Your prediction
Three data qubits · two parity ancillas · one ideal round

A parity result of 0 means the pair agrees; 1 means it differs.

q00
q10
q20
q0 ⊕ q1—q1 ⊕ q2—

The checks compare neighboring data bits; a dash means the syndrome has not yet been read.

A chosen X gate is part of the model circuit; no random hardware noise is being invented.

Logical bit 0. This demonstration protects a classical bit against one X error, not an unknown quantum state against all errors.

What did the HRL team actually demonstrate?

The paper includes repeated distance-3 and distance-5 repetition-code experiments, leakage-reduction operations, and a separate [[4,2,2]] quantum error-detecting code. This model implements none of their device-specific pulses or leakage dynamics.

For the quantum error-detection study, the authors report approximately 95% conditional logical-state fidelity after rejecting about 77% of shots, versus about 59% average fidelity when detection records were ignored. The accepted fraction matters alongside the quality of accepted results.

These are rounded, reported study summaries—not new QRL hardware runs or reconstructed raw counts. Post-selection is not the same as correcting every run.

Inspect the paper’s experiments and Figure 4 ↗

Checking voice availability…

Exact circuit probabilities

Theoretical probabilities for the selected step and basis.

Exact numerical values
Ideal probabilities at Prepare a data bit
OutcomeProbability
0001.0000000
0010.0000000
0100.0000000
0110.0000000
1000.0000000
1010.0000000
1100.0000000
1110.0000000

Fixed 0–100% scale · left digit is qubit 0

Z-basis measurement probabilities. A histogram is not the complete quantum state.

What this shows

Prepare a data bit

The encoded classical bit is 000 or 111. Checks q0⊕q1 and q1⊕q2 locate a single bit flip without directly reading the stored bit. Syndrome 10 indicates q0, 11 indicates q1, 01 indicates q2, and 00 means no single flip. A conditional X restores the encoded data.

Where this model stops

This is one ideal syndrome-extraction round for a classical logical bit with a deliberately injected X gate. It is not HRL’s exchange-only hardware, repeated distance-5 experiment, leakage-reduction sequence or [[4,2,2]] quantum error-detection experiment. It does not protect against arbitrary quantum errors or demonstrate fault tolerance.

Real QPU eligibility and source paper

The export uses three data qubits, two parity ancillas and conditional X. It requires validated dynamic-circuit hardware. No QPU counts or paper performance numbers are simulated here.

There is no hardware execution behind these plots. Exporting a circuit does not validate a device or create a new hardware result.

HRL Quantum Team and Collaborators · A digitally controlled silicon quantum processing unit ↗

Nature · July 29, 2026; Krishna Choudhary is a co-author

Predict · 1 of 6

Inspect the circuit

Export the selected step and settings for Qiskit. No job is submitted.

View Qiskit code
# Quantum Reality Lab — Recover a logical bit
# Reduced-scale repetition-code model; not a reproduction of the cited hardware experiment.
# Source: https://www.nature.com/articles/s41586-026-10754-7
# Display convention: qubit 0 is the leftmost display digit.
# Qiskit single-register count keys are reversed before display.
from qiskit import QuantumCircuit, QuantumRegister, ClassicalRegister
from math import pi

q = QuantumRegister(5, 'q')
syndrome = ClassicalRegister(2, 'syndrome')
data = ClassicalRegister(3, 'data')
qc = QuantumCircuit(q, syndrome, data)
# q0,q1,q2=data; q3,q4=parity ancillas. One round, no leakage model.
qc.measure([0, 1, 2], data)
# Parse separate 'data syndrome' register blocks explicitly.
# Reverse the three-bit data block for qubit-0-left display.
# This stores a classical logical bit, not an arbitrary protected quantum state.

print(qc.draw())