News for hackers. Stuff that matters.
Re:Cerebras-GPT vs. LLaMA AI Model Performance Comparison by wongarsu on Wednesday March 29, 2023 @07:18PM (#35365311)
even worse than sequential things not being sequential: 10000 is tokenized as one token (token 33028), while 10050 is two tokens, 100 (token 3064) and 50 (token 1120)

There are no comments on this comment yet.