[BUG] default.qubit.legacy returns wrong jacobian with qml.QubitUnitary+requires_grad=False

[vincentmr]

Expected behavior

I expect adjoint to return the same results as parameter-shift for the Jacobian of the circuit below.

[-0.13893233 -0.31571638 -0.62490128]

Actual behavior

It returns a different Jacobian

[ 0.         -0.13893233 -0.31571638]

Additional information

No response

Source code

import pennylane as qml
import pennylane.numpy as np
n_wires = 3
n_targets = 2

dev = qml.device("default.qubit.legacy", wires=n_wires, )

par = np.arange(1, n_wires+1) / 3

def circuit(x):
    for i in range(n_wires):
        qml.RY(x[i], wires=i)
    qml.QubitUnitary(np.eye(4, requires_grad=False), wires=range(n_targets))
    return qml.expval(qml.PauliZ(0)@qml.PauliZ(1)@qml.PauliZ(2))

circ = qml.QNode(circuit, dev, diff_method="adjoint")
print(qml.jacobian(circ)(par))
circ = qml.QNode(circuit, dev, diff_method="parameter-shift")
print(qml.jacobian(circ)(par))

Tracebacks

No response

System information

Name: PennyLane
Version: 0.34.0.dev0
Summary: PennyLane is a Python quantum machine learning library by Xanadu Inc.
Home-page: https://github.com/PennyLaneAI/pennylane
Author: 
Author-email: 
License: Apache License 2.0
Location: /global/u2/v/vincentm/repos/pennylane-lightning/venv/lib/python3.11/site-packages
Requires: appdirs, autograd, autoray, cachetools, networkx, numpy, pennylane-lightning, requests, rustworkx, scipy, semantic-version, toml, typing-extensions
Required-by: PennyLane-Lightning

Platform info:           Linux-5.14.21-150400.24.81_12.0.86-cray_shasta_c-x86_64-with-glibc2.31
Python version:          3.11.4
Numpy version:           1.26.1
Scipy version:           1.11.3
Installed devices:
- lightning.qubit (PennyLane-Lightning-0.34.0.dev5)
- default.gaussian (PennyLane-0.34.0.dev0)
- default.mixed (PennyLane-0.34.0.dev0)
- default.qubit (PennyLane-0.34.0.dev0)
- default.qubit.autograd (PennyLane-0.34.0.dev0)
- default.qubit.jax (PennyLane-0.34.0.dev0)
- default.qubit.legacy (PennyLane-0.34.0.dev0)
- default.qubit.tf (PennyLane-0.34.0.dev0)
- default.qubit.torch (PennyLane-0.34.0.dev0)
- default.qutrit (PennyLane-0.34.0.dev0)
- null.qubit (PennyLane-0.34.0.dev0)

Existing GitHub issues

• I have searched existing GitHub issues to make sure the issue does not already exist.

[vincentmr]

#4820

[Alex-Preciado]

This issue is now closed by PR #4820