diff --git a/examples/qasper.py b/examples/qasper.py
index 3bbe844..1cc34ee 100644
--- a/examples/qasper.py
+++ b/examples/qasper.py
@@ -45,21 +45,23 @@ def transform(self, data: TransformElementType) -> TransformElementType:
 def main():
     dataset = load_dataset("qasper", split="validation")
 
+    template = """
+    {{title}}{{abs_sep}}
+    {{abstract}}{{abs_sep}}
+    {% for i in range(full_text['section_name'] | length) %}
+    {{full_text['section_name'][i]}}{{title_sep}}
+    {% for paragraph in full_text['paragraphs'][i] %}
+    {{paragraph}}{{para_sep}}
+    {% endfor %}
+    {{sec_sep}}
+    {% endfor %}
+    """
+
     pipeline = (
         # concatenate the full text into a single string; use
         # title_sep, para_sep, sec_sep, and abs_sep to manage separators
         sm.JinjaMapper(
-            jinja=(
-                "{{title}}{{abs_sep}}"
-                "{{abstract}}{{abs_sep}}"
-                "{% for i in range(full_text['section_name'] | length) %}"
-                "{{full_text['section_name'][i]}}{{title_sep}}"
-                "{% for paragraph in full_text['paragraphs'][i] %}"
-                "{{paragraph}}{{para_sep}}"
-                "{% endfor %}"
-                "{{sec_sep}}"
-                "{% endfor %}"
-            ),
+            jinja=template,
             source_field_name="context",
             extra_variables={
                 "title_sep": "\n",
diff --git a/src/smashed/mappers/promptsource.py b/src/smashed/mappers/promptsource.py
index 38cadc9..0a77f04 100644
--- a/src/smashed/mappers/promptsource.py
+++ b/src/smashed/mappers/promptsource.py
@@ -11,7 +11,6 @@
     Sequence,
     Set,
     Tuple,
-    Type,
     Union,
     cast,
 )
@@ -49,13 +48,11 @@ class JinjaEnvironment:
     _env: Optional["Environment"] = None
 
     @classmethod
-    def env(cls, loader: Optional[Type["BaseLoader"]] = None) -> "Environment":
+    def env(cls, loader: Optional["BaseLoader"] = None) -> "Environment":
         if cls._env is not None:
             return cls._env
 
-        cls._env = Environment(
-            loader=(loader or BaseLoader)  # pyright: ignore
-        )
+        cls._env = Environment(loader=loader)
         return cls._env
 
     @classmethod
@@ -143,13 +140,16 @@ def approx_input_fields(self) -> Tuple[Set[str], ...]:
         since we can't parse out cases where for loops or if statements
         are used, nor cases where members of a variable are accessed.
         """
+
+        # we compute variables first because some might
+        all_variables = set(
+            field
+            for field in self.get_vars_from_txt(self.template)
+            if field not in self.extra_vars
+        )
         return tuple(
-            set(
-                field
-                for field in self.get_vars_from_txt(t)
-                if field not in self.extra_vars
-            )
-            for t in self.template.split("|||")
+            {v for v in all_variables if v in fragment}
+            for fragment in self.template.split("|||")
         )
 
     @property