diff --git a/tuw_nlp/sem/hrg/Documentation.md b/tuw_nlp/sem/hrg/Documentation.md
index 1b2badb..76b25d8 100644
--- a/tuw_nlp/sem/hrg/Documentation.md
+++ b/tuw_nlp/sem/hrg/Documentation.md
@@ -63,7 +63,7 @@ python steps/predict/predict.py -d $DATA_DIR -c pipeline/config/predict_100.json
 Once all sentences are predicted, we [merge](steps/predict/merge.py) them into one json per model.
 
 ```bash
- python steps/predict/merge.py -d $DATA_DIR -c pipeline/config/merge_100.json
+python steps/predict/merge.py -d $DATA_DIR -c pipeline/config/merge_100.json
 ```
 
 #### Run the whole predict pipeline on dev
@@ -81,10 +81,20 @@ python pipeline/pipeline.py -d $DATA_DIR -c pipeline/config/pipeline_dev_300.jso
 
 ### Create a random predictions for comparison
 
-We implement a [random extractor](random/random_extractor.py) that uses the [artefacts](random/train_stat) of the training dataset (distribution of the number of extractions per sentence, and distribution of labels per length of the sentence) and assures that the predicate is a verb.  
+We implement a [random extractor](steps/random/random_extractor.py) that uses the [artefacts](pipeline/output/artefacts) of the training dataset (distribution of the number of extractions per sentence, and distribution of labels per length of the sentence) and assures that the predicate is a verb.  
 
 ```bash
-# TBD
+# Extract artefacts
+python steps/random/artefacts.py -d $DATA_DIR -c pipeline/config/artefacts_train.json
+
+# Get random extractions
+python steps/random/random_extractor.py -d $DATA_DIR -c pipeline/config/random_dev.json
+
+# Merge the extractions
+python steps/predict/merge.py -d $DATA_DIR -c pipeline/config/merge_dev_random.json
+
+# Or run as a pipeline
+python pipeline/pipeline.py -d $DATA_DIR -c pipeline/config/pipeline_dev_random.json
 ```
 
 ### Evaluate the predictions
diff --git a/tuw_nlp/sem/hrg/pipeline/config/merge_dev_random.json b/tuw_nlp/sem/hrg/pipeline/config/merge_dev_random.json
new file mode 100644
index 0000000..fd32f12
--- /dev/null
+++ b/tuw_nlp/sem/hrg/pipeline/config/merge_dev_random.json
@@ -0,0 +1,14 @@
+{
+    "in_dir": "dev_random",
+    "k": 10,
+    "bolinas_chart_filters": 
+    [
+        "boa",
+        "argidx"
+    ],
+    "postprocess": 
+    [
+        ""
+    ],
+    "out_dir": "dev_extractions"
+}
\ No newline at end of file
diff --git a/tuw_nlp/sem/hrg/pipeline/config/pipeline_dev_random.json b/tuw_nlp/sem/hrg/pipeline/config/pipeline_dev_random.json
new file mode 100644
index 0000000..8f5fbd4
--- /dev/null
+++ b/tuw_nlp/sem/hrg/pipeline/config/pipeline_dev_random.json
@@ -0,0 +1,20 @@
+{
+    "steps":
+    [
+        {
+            "step_name": "artefacts",
+            "script_name": "artefacts",
+            "config": "artefacts_train.json"
+        },
+        {
+            "step_name": "random",
+            "script_name": "random",
+            "config": "random_dev.json"
+        },
+        {
+            "step_name": "merge_random",
+            "script_name": "merge",
+            "config": "merge_dev_random.json"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/tuw_nlp/sem/hrg/pipeline/config/random_dev.json b/tuw_nlp/sem/hrg/pipeline/config/random_dev.json
index 4623ffe..0e39e75 100644
--- a/tuw_nlp/sem/hrg/pipeline/config/random_dev.json
+++ b/tuw_nlp/sem/hrg/pipeline/config/random_dev.json
@@ -1,5 +1,7 @@
 {
+    "in_dir": "dev_preproc",
     "out_dir": "dev_random",
+    "artefact_prefix": "artefacts_train",
     "models": 
     [
         "boa",