Improve 0.4 release note prelude formatting

releasenotes/notes/0.4/drop_python_3_8-dffbed172a8cccf1.yaml

@@ -1,6 +1,4 @@
 ---
-prelude: >
-    Support for Python 3.8 has been dropped.
 upgrade:
   - |
     Python 3.8 support was dropped as Python 3.8 reached End Of Life

releasenotes/notes/0.4/numpy_2.0_fix-c29681db4874eee8.yaml

@@ -1,6 +1,6 @@
 ---
 fixes:
   - |
-    Compatibility fix for scipy 1.14 version. csr_matrix.H had been deprecated
+    Compatibility fix for scipy 1.14 version. ``csr_matrix.H`` had been deprecated
     and was removed. The :class:`.NumPyEigensolver` algorithm, which
     had been using that, has been updated to use an alternative.

releasenotes/notes/0.4/prelude_0.4_release-a9d23b9c63550bd2.yaml

@@ -2,10 +2,39 @@
 prelude: >
     This release compiles a series of important updates to the package, including:
 
-    - dropped support for Python 3.8 and added support for Python 3.13
-    - added support for PrimitivesV2
-    - addressed deprecations from the 1.x cycle
-
-    These changes allow for qiskit-algorithms to now support 
+        * dropped support for Python 3.8 and added support for Python 3.13
+        * added support for ``PrimitivesV2``
+        * addressed deprecations from the Qiskit 1.x cycle
+        
+    These changes allow for ``qiskit-algorithms`` to now support 
     `Qiskit 2.x <https://docs.quantum.ibm.com/api/qiskit/release-notes/2.0>`__
-    while maintaining support for the 1.x cycle.
+    while maintaining support for the 1.x cycle.
+
+    In further detail:
+
+        Following up on the deprecation of V1 primitives in Qiskit 1.X and their
+        subsequent removal in Qiskit 2.X, this release of ``qiskit-algorithms``
+        drops their support in favor of the V2 primitives. You can read about
+        how to modify your imports in your code in the
+        `Qiskit migration guide to the V2 primitives <https://quantum.cloud.ibm.com/docs/en/migration-guides/v2-primitives>__.
+        Since V2 primitives may require users to transpile their circuit, it is
+        now possible for the users to provide the classes that define their own
+        :class:`~qiskit.circuit.QuantumCircuit` with a ``Transpiler``
+        along with some options. A ``Transpiler`` is any object having a ``run``
+        method that can take as input a
+        :class:`~qiskit.circuit.QuantumCircuit` or a list thereof and
+        additional options and returns the transpiled version of its input(s)
+        according to the provided options.
+
+        Following up on the deprecation of ``BlueprintCircuit`` in the 2.1 version of Qiskit, most
+        tests have been updated to use the replacements functions instead. These circuits are still
+        supported by ``qiskit-algorithms``, though their use is now deprecated, and their support will
+        be removed when the oldest supported Qiskit version is 3.0.
+
+        The :class:`.AmplificationProblem` and :class:`.Grover` classes now support being
+        passed :class:`~qiskit.circuit.library.Gate` objects to support the use of
+        :class:`~qiskit.circuit.library.PhaseOracleGate` in addition to
+        :class:`~qiskit.circuit.library.PhaseOracle`.
+
+        The :class:`.AdaptVQE` class now supports a new way to specify its ansatz, following up on the
+        deprecation of :class:`~qiskit.circuit.library.EvolvedOperatorAnsatz`.

releasenotes/notes/0.3/raise-error-incorrect-bounds-passing-c878f155221fa8f4.yaml → releasenotes/notes/0.4/raise-error-incorrect-bounds-passing-c878f155221fa8f4.yaml

@@ -1,6 +1,6 @@
 ---
 fixes:
   - |
-    Resolved the issue with multiply-defined Scipy bounds in Optimizers. In line with Scipy, now only the ``minimize()``
+    Resolved the issue with multiply-defined Scipy bounds in :mod:`qiskit_algorithms.optimizers`. In line with Scipy, now only the ``minimize()``
     method supports the ``bounds`` keyword. An error is raised when trying to pass ``bounds`` in the Optimizer constructor
     via ``kwargs`` or ``options``.

releasenotes/notes/0.4/replaced-v1-primitives-by-v2-ones-897d9610351bd54b.yaml

@@ -1,18 +1,4 @@
 ---
-prelude: >
-    Following up on the deprecation of V1 primitives in Qiskit 1.X and their
-    subsequent removal in Qiskit 2.X, this release of ``qiskit-algorithms``
-    drops their support in favor of the V2 primitives. You can read about
-    how to modify your imports in your code in the
-    [Qiskit migration guide to the V2 primitives](https://quantum.cloud.ibm.com/docs/en/migration-guides/v2-primitives).
-    Since V2 primitives may require users to transpile their circuit, it is
-    now possible for the users to provide the classes that define their own
-    :class:`~qiskit.circuit.QuantumCircuit` with a ``Transpiler``
-    along with some options. A ``Transpiler`` is any object having a ``run``
-    method that can take as input a
-    :class:`~qiskit.circuit.QuantumCircuit` or a list thereof and
-    additional options and returns the transpiled version of its input(s)
-    according to the provided options.
 features:
   - |
     Some classes, notably those that create their own
@@ -23,7 +9,7 @@ features:
     the transpiled version of its input(s) according to the provided options.
     Passing a ``Transpiler`` to such a class is done via the `transpiler`
     keyword-only argument at instantiation, and additional options passed to
-    its `run` method can be passed using the `transpiler_options` keyword-only
+    its ``run`` method can be passed using the ``transpiler_options`` keyword-only
     argument at instantiation. The classes that support this feature are
 
       - :class:`.Grover`
@@ -45,7 +31,7 @@ features:
     to those above, these classes don't create their own
     :class:`qiskit.circuit.QuantumCircuit`, which is instead provided by the user.
     As such, transpiling the circuits prior to passing them to theses classes
-    has the same effect as providing these classes with a `Transpiler`. These classes are
+    has the same effect as providing these classes with a ``Transpiler``. These classes are
 
       - :class:`.FiniteDiffEstimatorGradient`
       - :class:`.FiniteDiffSamplerGradient`

releasenotes/notes/0.4/sampler_as_positional_arg_for_qnspsa-2efca170947f3cd7.yaml

@@ -1,7 +1,7 @@
 ---
 upgrade:
   - |
-    The list of arguments for `QNSPSA.get_fidelity()` has been updated to
-    remove the keyword enforcement for `sampler` so also allowing it now
-    as positional. Additionally the `sampler` must be explicitly specified, as a
+    The list of arguments for :meth:`.QNSPSA.get_fidelity()` has been updated to
+    remove the keyword enforcement for ``sampler`` so also allowing it now
+    as positional. Additionally the ``sampler`` must be explicitly specified, as a
     valid object, as it no longer supports `None` nor defaults to that.