π“Ÿ host1.bitchan.one

𓀐 @-wtHrmgM

⏱ 1w seenβ›“ 0πŸ‘ 0β§– 1

filter [-]

languages:onother

Is it just chinese open models who have reasoning in caveman speak? Or do closed sota ones do that too? Is it really useful?⏱ 1w↩ 1
They should natively think in Chinese. That feels like it would be very token-efficient.⏱ 1w↩