diff --git a/scripts/pre_train/ultra-fineweb-l3-en-qa-synthetic.py b/scripts/pre_train/ultra-fineweb-l3-en-qa-synthetic.py
new file mode 100644
index 0000000..eab46eb
--- /dev/null
+++ b/scripts/pre_train/ultra-fineweb-l3-en-qa-synthetic.py
@@ -0,0 +1,20 @@
+from datasets import load_dataset
+from pipeline import export_dataset
+
+
+def process_func(input_dict: dict):
+    return {"text": input_dict["content"]}
+
+
+if __name__ == "__main__":
+    dataset = load_dataset(
+        "openbmb/Ultra-FineWeb-L3",
+        "Ultra-FineWeb-L3-en-QA-Synthetic",
+        split="train",
+    )
+    export_dataset(
+        dataset=dataset,
+        output_dir="./dataset",
+        output_prefix="ultra-fineweb-l3-en-qa-synthetic-pretrain",
+        process_func=process_func,
+    )
diff --git a/scripts/pre_train/ultra-fineweb-l3-zh-qa-synthetic.py b/scripts/pre_train/ultra-fineweb-l3-zh-qa-synthetic.py
new file mode 100644
index 0000000..e7b005f
--- /dev/null
+++ b/scripts/pre_train/ultra-fineweb-l3-zh-qa-synthetic.py
@@ -0,0 +1,20 @@
+from datasets import load_dataset
+from pipeline import export_dataset
+
+
+def process_func(input_dict: dict):
+    return {"text": input_dict["content"]}
+
+
+if __name__ == "__main__":
+    dataset = load_dataset(
+        "openbmb/Ultra-FineWeb-L3",
+        "Ultra-FineWeb-L3-zh-QA-Synthetic",
+        split="train",
+    )
+    export_dataset(
+        dataset=dataset,
+        output_dir="./dataset",
+        output_prefix="ultra-fineweb-l3-zh-qa-synthetic-pretrain",
+        process_func=process_func,
+    )