diff --git a/doc/README.md b/doc/README.md
index e90a3014d..aa53bd3d4 100644
--- a/doc/README.md
+++ b/doc/README.md
@@ -134,7 +134,7 @@ The script `../scripts/nb2doc/convert.sh` can be used:
 * to generate documents from new notebooks;
 * to ensure that any changes to the existing notebooks are reflected in the existing documents (for instance in a CI setting).
 
-The script must be run from the project root directory and requires [Pandoc](https://pandoc.org/) to be already installed. The latest supported version of Pandoc is 3.3
+The script must be run from the project root directory and requires [Pandoc](https://pandoc.org/) to be already installed. The latest supported version of Pandoc is 3.5
 
 ```bash
 ./scripts/nb2doc/convert.sh
diff --git a/doc/modules/ROOT/pages/tutorials/heterogeneous-node-classification-with-hashgnn.adoc b/doc/modules/ROOT/pages/tutorials/heterogeneous-node-classification-with-hashgnn.adoc
index 3f6ada73a..6ddf9fb2f 100644
--- a/doc/modules/ROOT/pages/tutorials/heterogeneous-node-classification-with-hashgnn.adoc
+++ b/doc/modules/ROOT/pages/tutorials/heterogeneous-node-classification-with-hashgnn.adoc
@@ -347,7 +347,7 @@ part with something like a
 https://neo4j.com/docs/graph-data-science-client/current/graph-object/#_projecting_a_graph_object[projection
 from a Neo4j database] to create a more realistic production workflow.
 
-== Comparison with other methods
+=== Comparison with other methods
 
 As mentioned we tried to mimic the setup of the benchmarks in the
 NeurIPS paper
@@ -382,7 +382,7 @@ descent compared to the deep learning models above - HashGNN is an
 unsupervised algorithm - HashGNN runs a lot faster (even without a GPU)
 and requires a lot less memory
 
-== Further learning
+=== Further learning
 
 To learn more about the topics covered in this notebook, please check
 out the following pages of the GDS manual: