diff --git a/wtpsplit/evaluation/intrinsic.py b/wtpsplit/evaluation/intrinsic.py
index 067cb8a8..d44844b0 100644
--- a/wtpsplit/evaluation/intrinsic.py
+++ b/wtpsplit/evaluation/intrinsic.py
@@ -267,14 +267,17 @@ def main(args):
                 "t": score_t,
                 "punct": score_punct,
             }
+            
+            if score_u is not None:
+                u_scores.append((score_u, lang_code))
+            if score_t is not None:
+                t_scores.append((score_t, lang_code))
+            if score_punct is not None:
+                punct_scores.append((score_punct, lang_code))
 
             # just for printing
             score_t = score_t or 0.0
             score_punct = score_punct or 0.0
-
-            u_scores.append((score_u, lang_code))
-            t_scores.append((score_t, lang_code))
-            punct_scores.append((score_punct, lang_code))
             print(f"{lang_code} {dataset_name} {score_u:.3f} {score_t:.3f} {score_punct:.3f}")
 
     # Compute statistics for each metric across all languages
diff --git a/wtpsplit/evaluation/time_intrinsic.py b/wtpsplit/evaluation/time_intrinsic.py
index 83f89021..0c8a0b89 100644
--- a/wtpsplit/evaluation/time_intrinsic.py
+++ b/wtpsplit/evaluation/time_intrinsic.py
@@ -39,6 +39,7 @@ def benchmark_strides(low_stride, high_stride, args):
                 "max_n_train_sentences": args.max_n_train_sentences,
             }
         )
+        print(results_data)
 
     return pd.DataFrame(results_data)
 
@@ -73,5 +74,5 @@ def benchmark_strides(low_stride, high_stride, args):
     # Optionally save df_results to a file
     # to csv
     df_results.to_csv(
-        f"timing_results_{args.model_path.replace('/','__')}_batch{args.batch_size}_b{args.block_size}+s{args.stride}_u{args.threshold}_AVG.csv"
+        f"timing_results_{args.model_path.replace('/','__')}_batch{args.batch_size}_b{args.block_size}+s{args.stride}_n{args.max_n_train_sentences}_u{args.threshold}_AVG.csv"
     )