A / REFERENCE
AThe baseline
Enter a preparation to see its gates.
NO CIRCUIT LOADED
◆ THE QUANTUM OBSERVATORY EXPERIMENT 001
Look beyond
the bars.
Two circuits can look the same in ordinary measurements. The difference might be hiding in their phase.
↓ SCROLL TO EXPLOREDESIGNED FOR DISCOVERY
01
PROCESSOR MODEL
◆ YOUR EXPERIMENT A / B COMPARISON
From circuit to insight.
Build your comparison, choose a measurement plan, and follow the evidence.
PHASESCOPE / QUANTUM CIRCUIT COMPARISON
One gate changed. Can the measurements tell?
Two Bell-state circuits on IBM Quantum hardware. The candidate adds a single Z gate to q₀.
STEP 01 / THE INPUTS
OPENQASM 2 · TWO QUBITSMeet your circuits.
Start with two preparations. PhaseScope adds the measurements later.
B / CANDIDATE
BThe comparison
Enter a preparation to see its gates.
NO CIRCUIT LOADED
Normalized gate changes
✳
A small change, a different question.
PhaseScope validates each preparation locally and draws the gates in source order.
STEP 02 / BEFORE YOU RUN
MEASUREMENT PLANA closer look takes nine views.
Measure both preparations in each local Pauli basis. Settings are labeled q₁ ⊗ q₀.
THE MEASUREMENT GRID
SELECT A BASIS TO INSPECTNine perspectives. One comparison.
ZZ
The ordinary measurement view
Measures both qubits in the computational basis. It can miss differences in relative phase.
STEP 03 / HARDWARE RUN
HARDWARE UNCHECKEDReady when you are.
Real hardware results are the foundation of every comparison.
IBM QUANTUM
Compile on a physical pair
Check access to the server’s configured processor and select a pair. Compilation verifies that every measurement stays on those qubits.
PERSISTED EXPERIMENTS
Run history
Job provenance
STEP 04 / THE EVIDENCE
NO COMPLETED HARDWARE RESULTSSee what the bars miss.
Inspect counts, shot-noise intervals, and reconstructed matrices from the selected run.
A / REFERENCE
B / CANDIDATE
These 95% intervals cover shot noise. Drift, correlated shots, and hardware systematics are outside them. An unresolved difference does not establish equality. Arrow keys move ZZ, XX, and YY.
◌No counts to display
Distributions will appear here after a hardware run.
00011011
How these counts become a correlation
Matching bits (00, 11) count as +1. Different bits (01, 10) count as −1. Divide their difference by the total shots.
Basis labels read q₁ ⊗ q₀. The rotations are for the logical circuits before hardware transpilation. IBM guide to Pauli measurements
REFERENCE A CANDIDATE BRETURNED SHOTS: —
Where did the circuits differ?
Each dot shows the measured change from A to B. The line shows its 95% shot-noise interval. If the interval includes zero, the difference is unresolved at this shot count.
How phase changes the measurements
Start with (|00⟩ + eiφ|11⟩) / √2. The dial shows the angle of the |11⟩ amplitude relative to |00⟩. The added Z gate moves the ideal state from 0° to 180°.
Predicted correlations
| Basis | Ideal value | Matching bits |
|---|---|---|
| ZZ | ||
| XX | ||
| YY | ||
| XY / YX |
In ZZ, the ideal probabilities of 00 and 11 are 50% each at every angle. XX and YY respond to cos φ. XY and YX respond to sin φ.
Measured endpoints · ibm_fez
These are fixed results for the two prepared circuits. The dial changes the ideal model above.
| Basis | Reference A | Candidate B |
|---|
Observable differences
95% family-level shot-noise intervals across 30 estimates. Drift, correlated shots, and hardware systematics are outside these bounds.
| Observable | Basis | A estimate [interval] | B estimate [interval] | B − A [interval] | Resolution |
|---|
RECONSTRUCTED DENSITY MATRICES
Look between the outcomes.
Estimated entries appear after all required measurements are complete.
The coherence changes sign
The real part of ⟨00|ρ|11⟩ equals (XX − YY) / 4 in this reconstruction.
A / REFERENCE
B / CANDIDATE
REFERENCE A / REAL
CANDIDATE B / REAL
B − A / REAL
−10+1
Select an entry to inspect its role. Basis order: |q₁ q₀⟩ = 00, 01, 10, 11.
STEP 05 / TAKE IT WITH YOU
EXPERIMENT PACKETKeep the evidence.
Preserve the details behind your comparison, not just the picture.
↗
NOT YET AVAILABLEYour experiment packet lives here.
Select a finished run to download its evidence packet.
Download experiment packet ↓Open read-only report ↗Verify the evidence locally
Extract the ZIP, install NumPy, then run python verify_evidence.py phasescope-evidence.zip against the downloaded file.
Published ZIP SHA-256
The verifier recomputes the analysis. IBM provenance requires an independent source.