diff --git a/convert.py b/convert.py
index e9b08d344f5bd..96b6c6f847e1c 100755
--- a/convert.py
+++ b/convert.py
@@ -366,44 +366,47 @@ def __init__(self, fname_tokenizer: Path, fname_added_tokens: Path | None) -> No
             added_tokens = {}
 
         vocab_size: int = self.sentencepiece_tokenizer.vocab_size()
-        expected_ids = list(range(vocab_size, vocab_size + len(added_tokens)))
-        actual_ids   = sorted(added_tokens.values())
-        if expected_ids != actual_ids:
-            raise Exception(f"Expected added token IDs to be sequential and start at {len(added_tokens)}; got {actual_ids}")
 
-        items = sorted(added_tokens.items(), key=lambda text_idx: text_idx[1])
-        self.added_tokens_list = [text for (text, idx) in items]
-        self.vocab_size_base: int = vocab_size
-        self.vocab_size: int = self.vocab_size_base + len(self.added_tokens_list)
-        self.fname_tokenizer = fname_tokenizer
-        self.fname_added_tokens = fname_added_tokens
+        new_tokens: dict[int, str] = {id: piece for piece, id in added_tokens.items() if id >= vocab_size}
+        expected_new_ids: list[int] = list(range(vocab_size, vocab_size + len(new_tokens)))
+        actual_new_ids: list[int]   = sorted(new_tokens.keys())
+
+        if expected_new_ids != actual_new_ids:
+            raise Exception(f"Expected new token IDs {expected_new_ids} to be sequential; got {actual_new_ids}")
+
+        # Token pieces that were added to the base vocabulary.
+        self.new_tokens_list: list[str] = [new_tokens[id] for id in actual_new_ids]
+        self.vocab_size_base: int       = vocab_size
+        self.vocab_size: int            = self.vocab_size_base + len(self.new_tokens_list)
+        self.fname_tokenizer            = fname_tokenizer
+        self.fname_added_tokens         = fname_added_tokens
 
     def sentencepiece_tokens(self) -> Iterable[tuple[bytes, float, gguf.TokenType]]:
         tokenizer = self.sentencepiece_tokenizer
-        for i in range(tokenizer.vocab_size()):
-            piece = tokenizer.id_to_piece(i)
+        for id in range(tokenizer.vocab_size()):
+            piece = tokenizer.id_to_piece(id)
             text: bytes = piece.encode("utf-8")
-            score: float = tokenizer.get_score(i)
+            score: float = tokenizer.get_score(id)
 
             toktype = gguf.TokenType.NORMAL
-            if tokenizer.is_unknown(i):
+            if tokenizer.is_unknown(id):
                 toktype = gguf.TokenType.UNKNOWN
-            if tokenizer.is_control(i):
+            if tokenizer.is_control(id):
                 toktype = gguf.TokenType.CONTROL
 
             # NOTE: I think added_tokens are user defined.
             # ref: https://github.com/google/sentencepiece/blob/master/src/sentencepiece_model.proto
             # if tokenizer.is_user_defined(i): toktype = gguf.TokenType.USER_DEFINED
 
-            if tokenizer.is_unused(i):
+            if tokenizer.is_unused(id):
                 toktype = gguf.TokenType.UNUSED
-            if tokenizer.is_byte(i):
+            if tokenizer.is_byte(id):
                 toktype = gguf.TokenType.BYTE
 
             yield text, score, toktype
 
     def added_tokens(self) -> Iterable[tuple[bytes, float, gguf.TokenType]]:
-        for text in self.added_tokens_list:
+        for text in self.new_tokens_list:
             score = -1000.0
             yield text.encode("utf-8"), score, gguf.TokenType.USER_DEFINED