Falcon-H1R: Pushing the Reasoning Frontiers with a Hybrid Model for Efficient Test-Time Scaling
Beginner Falcon LLM Team, Iheb Chaabane et al.Jan 5arXiv
Falcon-H1R is a small (7B) AI model that thinks really well without needing giant computers.
#Falcon-H1R#Hybrid Transformer-Mamba#Chain-of-Thought