#Grok trained for two months and already surpassed GPT3.5 in the MMLU benchmark! FWIW:
– Grok-0 was trained on 33B parameters.
– Grok-1 has better mathematical and reasoning abilities.
Grok-1 Surpasses GPT-3.5 on MMLU After Two Months
By
–
By
–
#Grok trained for two months and already surpassed GPT3.5 in the MMLU benchmark! FWIW:
– Grok-0 was trained on 33B parameters.
– Grok-1 has better mathematical and reasoning abilities.