LessWrong (Curated & Popular) cover image

"Recent LLMs can use filler tokens or problem repeats to improve (no-CoT) math performance" by ryan_greenblatt

LessWrong (Curated & Popular)

00:00

Repeating problem statements boosts performance

Ryan explains that repeating problems also raises no‑COT performance, often matching or exceeding filler effects.

Play episode from 01:06
Transcript

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app