diff --git a/finetune/encoding/input_encoder.py b/finetune/encoding/input_encoder.py
index eddb6266..86c901fa 100644
--- a/finetune/encoding/input_encoder.py
+++ b/finetune/encoding/input_encoder.py
@@ -320,7 +320,7 @@ def tokenize_context(context, encoded_output, config):
                     )
             if (
                 context_by_char_loc[current_char_loc][2]
-                and token.strip().strip("Â") not in context_by_char_loc[current_char_loc][2]
+                and token.strip().strip("Âĉ") not in context_by_char_loc[current_char_loc][2]
             ):
                 warnings.warn(
                     "subtoken: {} has matched up with the context for: {}".format(