Jim Fan (@drjimfan) 's Twitter Profile
Jim Fan

@drjimfan

NVIDIA Sr. Research Manager. Co-Lead of GR00T (Humanoid Robotics) & GEAR Lab. Solving Physical AI, one motor at a time. Stanford Ph.D. OpenAI's 1st intern.

ID: 1007413134

linkhttps://jimfan.me calendar_today12-12-2012 22:11:27

3,3K Tweet

302,302K Takipçi

3,3K Takip Edilen

Jim Fan (@drjimfan) 's Twitter Profile Photo
Jim Fan
@drjimfan

2 years ago

How to build *TruthGPT*? I listened to a talk by the legendary John Schulman. It's densely packed with lots of deep insight. Key takeaways: - Supervised finetuning (or behavior cloning) makes the model prone to hallucination, while RL mitigates it. - NLP is far from done! 1/🧵

How to build *TruthGPT*? I listened to a talk by the legendary <a href="/johnschulman2/">John Schulman</a>. It's densely packed with lots of deep insight. Key takeaways: - Supervised finetuning (or behavior cloning) makes the model prone to hallucination, while RL mitigates it. - NLP is far from done! 1/🧵