Teortaxesโถ๏ธ (DeepSeek ๆจ็น๐้็ฒ 2023 โ โ)
@teortaxesTex
We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization.
@deepseek_ai stan #1#, 2023โDeep Time
ยซCโest la guerre.ยป ยฎ1
๊ฐ์
September 2010
3.3K ํ๋ก์ ์ค 71.8K ํฌ
There are two hypotheses for the DeepSeek-V4's strange performance (as in, V4-Flash is about as good as we expected, but V4-Pro is disappointing given its scale):
1) failed pretrain
2) big difference in the RL/MOPD stage
Flash probably got multiple such iterations
๋ ๋ณด๊ธฐ
