Maxime Labonne (@maximelabonne) 's Twitter Profile
Maxime Labonne

@maximelabonne

Head of Post-Training @LiquidAI_
📝 Blog: mlabonne.github.io/blog
💻 GitHub: github.com/mlabonne
🤗 HF: huggingface.co/mlabonne

ID: 923114064460558336

calendar_today25-10-2017 09:08:38

2,2K Tweet

21,21K Takipçi

521 Takip Edilen

Maxime Labonne (@maximelabonne) 's Twitter Profile Photo
Maxime Labonne
@maximelabonne

7 months ago

New Re-Wild instruction dataset outperforms strong baselines like Tulu 3. The authors re-generated answer from WildChat-1M with different open-weight models. Lots of lessons in this paper.

New Re-Wild instruction dataset outperforms strong baselines like Tulu 3. The authors re-generated answer from WildChat-1M with different open-weight models. Lots of lessons in this paper.