Fix test_cuda using nonexistent mlm_tokens fixture

The test_cuda function referenced a fixture named mlm_tokens which was
never defined, causing a fixture-not-found error. Changed to use the
existing tokens fixture which provides the same MLM-style tokenized input.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: brendanlong

tests/acceptance/test_hooked_encoder.py

@@ -222,6 +222,6 @@ def test_input_list_of_strings_mlm(our_bert, huggingface_bert, tokenizer):
 
 
 @pytest.mark.skipif(not torch.cuda.is_available(), reason="Requires a CUDA device")
-def test_cuda(mlm_tokens):
+def test_cuda(tokens):
     model = HookedEncoder.from_pretrained(MODEL_NAME)
-    model(mlm_tokens)
+    model(tokens)