Dean Ball says distillation likely explains more of DeepSeek's parity with OpenAI than he thought
OpenAI policy researcher Dean Ball says distillation plays a bigger role in Chinese labs' close-behind performance than he realized, revising his estimate from ~30% to ~50% after seeing what one can within OpenAI. He notes DeepSeek shares Astra's terse reasoning style, and argues much practical distillation is adversarial chain-of-thought extraction at scale, which he calls an "attack" and says is very hard to detect—training classifiers on extracted CoT has a "crazy high" false positive rate.
All posts · 9, most important first
- Distillation has always been part of the explanation for why China is so close behind the U.S. in model benchmark performance. But one of my updates after joining openai is that distillation probably plays a bigger role in the explanation than I realized.very curious that DeepSeek now has the same terse reasoning style as Astra must be a coincidence open on X7 likes · 2 reposts
- I would have guessed distillation explained ~30% of Chinese close-behindness before seeing what one can see within openai, and after that, more like 50%.2 likes · 1 reposts
- Right, but my point is that a big part of distillation in practice involves people stealing cots. It’s not just training on the normal outputs of the model. This is one of the reasons we tend to think of them as “attacks.”1 likes · 0 reposts
- much much harder to train classifiers on this than you’d think. false positive rate is crazy high.1 likes · 0 reposts
- To first approximation distillation is just getting access to the full CoT imo, and the Chinese labs have teams to build methods of extraction2 likes · 0 reposts
- well, there are limits to what I can say here, but I’d just note there’s space between “useful amount of extracted cot” and “full cot.” Generally I do think researchers and policy people mean quite different things by “distillation.”2 likes · 0 reposts
- I haven’t really shared this publicly because people will next-token complete the obvious cope “buhhhh you are just saying that because derrrr” replies but it really is true, regardless of whether you think it’s good or not.4 likes · 0 reposts
- we def do not “know” that. the reasoning traces are extracted at huge scale.1 likes · 0 reposts
- Elie, do you mind if I qt some of our earlier exchange to highlight this semantic distinction about “distillation as research method” and “distillation as adversarial cot extraction”? think it is a useful distinction to elaborate but don’t want to put you on blast1 likes · 0 reposts
@deanwball policy · 
@natolambert research ·