Pawel Garbacki (@pawelg) 's Twitter Profile
Pawel Garbacki

@pawelg

GenAI researcher, @FireworksAI_HQ co-founder, ex-Meta, ex-Google

ID: 30014374

calendar_today09-04-2009 16:24:54

35 Tweet

66 Takipçi

417 Takip Edilen

Pawel Garbacki (@pawelg) 's Twitter Profile Photo
Pawel Garbacki
@pawelg

7 months ago

Check out our latest blog post explaining how GRPO (Group Relative Policy Optimization), employed by models like DeepSeek R1, helps models learn effectively without the heavy lifting of value networks or massive supervised datasets: * By skipping a standalone Value Model, GRPO