6/11/2026 at 2:14:18 PM
> Barely amortising at the bottom. At small batch each new token added to the batch tends to activate fresh expertsWhether this is true depends on what you mean by small. In general, AIUI you don't need more than a handful of experts to get a meaningful probability of overlap. DeepSeek V4 Pro is an exceptionally sparse model and even there you start to get meaningful overlap for a batch size of 5 or more. Moreover, in general you can think of the average amount of activated experts for a batch of size b as being n(1 - (1 - k/n)^b) where k is the number of active and n of total experts. For DeepSeek V4, k=6 and n is 256 for Flash, 384 for Pro. (The sampling is repeated per layer, not just per token.)
by zozbot234
6/11/2026 at 3:27:17 PM
https://fergusfinn.com/blog/economics-of-speculative-decodin...good point tho - plus for Deepseek the shared expert increases the overlap slightly
by somnial
6/11/2026 at 2:57:50 PM
The article includes that formula too and takes the overlap into account in its calculations.by yorwba
6/11/2026 at 3:36:37 PM
True but OP says that there is a meaningful "knee" at b=n/k (about 43 for DeepSeek V4 Flash) and I'm not sure that's all that relevant. If anything, it might be a bit more meaningful to highlight the point where on average half the experts are covered, which is coincidentally around 43 for Pro and 30 for Flash. Since that ought to be approximately where the variance in that expectation is maximized.by zozbot234