Aayush Karan's banner

Aayush Karan

@aakaran31 • 1,662 subscribers

PhD student @harvard and @nvidia | Algorithmic insights for generative machine learning | @PDSoros 2024 | Prev @GoogleDeepMind, @citsecurities, @apple

Shorts

We found a new way to get language models to reason. 🤯 No RL, no training, no verifiers, no prompting. ❌ With better sampling, base models can achieve single-shot reasoning on par with (or better than!) GRPO while avoiding its characteristic loss in generation diversity.

We found a new way to get language models to reason. 🤯 No RL, no training, no verifiers, no prompting. ❌ With better sampling, base models can achieve single-shot reasoning on par with (or better than!) GRPO while avoiding its characteristic loss in generation diversity.

277,680 views