Comment by lyu07282

No it just looks at exact character sequences, try it out yourself here: https://platform.openai.com/tokenizer