Xander Chin's banner
Xander Chin's profile picture

Xander Chin

@XanderChin3,643 subscribers

inference @nvidia | eng @westernu @schulichleaders | building and learning for fun

Shorts

hand-controlled boids

hand-controlled boids

739,199 次观看

as an intro to mechanistic interpretability, i decided to look into the formation of induction heads, which are circuits that allow LLMs to perform in-context learning by searching for previous occurrences of a sequence to predict the next token to form these circuits i trained attention-only transformers to repeat varying sequence lengths of random tokens. by randomizing the sequence length, i prevented the models from relying on rote memorization, forcing them to instead develop a generalizable circuit during this dive, i recorded some really cool findings and saw some interesting visual patterns emerging:

as an intro to mechanistic interpretability, i decided to look into the formation of induction heads, which are circuits that allow LLMs to perform in-context learning by searching for previous occurrences of a sequence to predict the next token to form these circuits i trained attention-only transformers to repeat varying sequence lengths of random tokens. by randomizing the sequence length, i prevented the models from relying on rote memorization, forcing them to instead develop a generalizable circuit during this dive, i recorded some really cool findings and saw some interesting visual patterns emerging:

26,486 次观看

Videos

没有更多内容可加载