MIT EI seminar, Hyung Won Chung from OpenAI. "Don't teach. Incentivize."

Рет қаралды 24,316

Hyung Won Chung

Күн бұрын

Пікірлер: 36

@hengfun Ай бұрын

"teach him taste of fish and make him hungry"🤣

@kevon217 Ай бұрын

Love that extra layer to that analogy.

@patrickchristianmagtaan5511 Ай бұрын

😂

@buoyrina9669 Ай бұрын

I enjoyed every bit of this talk.

@BuddhaMedam 15 күн бұрын

Thank you very much for this goldly video sir .Was Definitely worth spending 35 mins on it .

@JimSlattery Ай бұрын

Thanks for the insightful talk! I love the clarity at 18:50 of seeing the LLM going through training, with so many skills implicitly demanded by the next token prediction task.

@vrushankdesai715 Ай бұрын

such an incredibly information dense talk, thank you!

@wwkk4964 Ай бұрын

One of the best talks on KZbin right now!

@JimSlattery Ай бұрын

I like your concept of scaling: 1) identify the modeling assumption or inductive biases that bottlenecks further scaling 2) replace it with a more scalable one. Example: letting the model learn it's own representations is a more available approach.

@KeldonLee Ай бұрын

very insightful, thanks.

@JohnSeong Ай бұрын

좋은 강의 무료로 올려주셔서 감사합니다!

@windmaple Ай бұрын

Somehow Hyung Won Chung's talk is always very abstract and purely at a meta level. He doesn't talk about specific LLM techniques or anything like that, but goes all into the fundamental intuition behind the scaling 👍

@swyxTV Ай бұрын

probably cause if he would get into trouble otherwise haha

@sughoshkaushik7261 Ай бұрын

This talk is gold

@honon-cs2wl Ай бұрын

형원님, 정말 멋지십니다. 구독하고 항상 응원할게요!! 앞으로도 많은 영상 부탁드려요

@YucheolChoi Ай бұрын

멋지십니다 형원님!!😊

@chenjus Ай бұрын

"No amount of bananas can incentivize monkeys to do mathematical reasoning" lol

@SONJOGYO Ай бұрын

좋은 강의 감사합니다. 끊임없는 배움의 해체(unlearning)에 대해 이야기하신 것이 세상을 보는 눈을 트이게 한 느낌이네요. 잘못된 공리를 바탕으로 구축된 직관과 아이디어가 해체되어야 한다는 이야기를 크게 생각해본 적이 없었으니까요.

@JimSlattery Ай бұрын

Great point: The Bitter Lesson article is the single most important writing in the field of AI. 😳

@saraswathiserene8142 19 күн бұрын

Which article? by Rich Sutton? Thankyou.

@wangqis Ай бұрын

谢谢这个分享~

@tiendatnguyen7527 Ай бұрын

thank you for sharing!

@andrewthomas9433 Ай бұрын

referenced talk: John Schulman - Reinforcement Learning from Human Feedback: Progress and Challenges kzbin.info/www/bejne/npnMfaprhsSIfMk

@조바이든-r6r Ай бұрын

멋있어요! 일이 많아서 힘드실거라고 생각이 듭니다. 건강도 챙기시길 바래요

@-mwolf Ай бұрын

yep - we should follow this principle in architecture too!

@roro5179 Ай бұрын

my goat

@optimaiz Ай бұрын

cool, thx u for sharing

@hsuai6584 Ай бұрын

It is suprised to know that you majored in mechanical engineering of your phd degree. How can you make such a big move?

@ThomasBrown2 Ай бұрын

@madrooky1398 Ай бұрын

"MIT EI seminar" reads like "With egg seminar" when you are German, super weird title.... Is it about breakfast? 😂

@DistortedV12 Ай бұрын

No Q/A?

@deter3 Ай бұрын

very shallow perspective .

@jashvira9487 Ай бұрын

Propagandic