[BUG] FABLE template returns incorrect value in JIT mode

[willjmax]

Expected behavior

The value returned by the FABLE template with JIT enabled should be approximately equal to the value without JIT enabled.

Actual behavior

The values are different.

Additional information

No response

Source code

def fable(input_matrix):
            qml.FABLE(input_matrix, wires=range(5), tol=0)
            return qml.expval(qml.PauliZ(wires=0))

        input_matrix = np.array(
            [
                [-0.5, -0.4, 0.6, 0.7],
                [0.9, 0.9, 0.8, 0.9],
                [0.8, 0.7, 0.9, 0.8],
                [0.9, 0.7, 0.8, 0.3],
            ]
        )

        device = qml.device("default.qubit", wires=5)
        interpreted_fn = qml.QNode(fable, device)
        jitted_fn = jax.jit(interpreted_fn)
        assert np.allclose(jitted_fn(input_matrix), interpreted_fn(input_matrix))

Tracebacks

No response

System information

Name: PennyLane
Version: 0.39.0.dev11
Summary: PennyLane is a cross-platform Python library for quantum computing, quantum machine learning, and quantum chemistry. Train a quantum computer the same way as a neural network.
Home-page: https://github.com/PennyLaneAI/pennylane
Author: 
Author-email: 
License: Apache License 2.0
Location: /home/william.maxwell/pennylane/pl-venv/lib/python3.9/site-packages
Editable project location: /home/william.maxwell/pennylane/pennylane
Requires: appdirs, autograd, autoray, cachetools, networkx, numpy, packaging, pennylane-lightning, requests, rustworkx, scipy, toml, typing-extensions
Required-by: amazon-braket-pennylane-plugin, PennyLane-Catalyst, PennyLane_Lightning, PennyLane_Lightning_Kokkos

Platform info:           Linux-6.8.0-40-generic-x86_64-with-glibc2.35
Python version:          3.9.19
Numpy version:           1.26.4
Scipy version:           1.12.0
Installed devices:
- default.clifford (PennyLane-0.39.0.dev11)
- default.gaussian (PennyLane-0.39.0.dev11)
- default.mixed (PennyLane-0.39.0.dev11)
- default.qubit (PennyLane-0.39.0.dev11)
- default.qubit.autograd (PennyLane-0.39.0.dev11)
- default.qubit.jax (PennyLane-0.39.0.dev11)
- default.qubit.legacy (PennyLane-0.39.0.dev11)
- default.qutrit (PennyLane-0.39.0.dev11)
- default.qutrit.mixed (PennyLane-0.39.0.dev11)
- default.tensor (PennyLane-0.39.0.dev11)
- null.qubit (PennyLane-0.39.0.dev11)
- lightning.qubit (PennyLane-Lightning-0.38.0)
- lightning.kokkos (PennyLane-Lightning-Kokkos-0.38.0)
- nvidia.custatevec (PennyLane-Catalyst-0.9.0.dev9)
- nvidia.cutensornet (PennyLane-Catalyst-0.9.0.dev9)
- oqc.cloud (PennyLane-Catalyst-0.9.0.dev9)
- softwareq.qpp (PennyLane-Catalyst-0.9.0.dev9)
- braket.aws.ahs (amazon-braket-pennylane-plugin-1.29.0)
- braket.aws.qubit (amazon-braket-pennylane-plugin-1.29.0)
- braket.local.ahs (amazon-braket-pennylane-plugin-1.29.0)
- braket.local.qubit (amazon-braket-pennylane-plugin-1.29.0)

Existing GitHub issues

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

[willjmax]

Fixed in #6263