s21mind/HexaMind-Llama-3.1-8B-v25-Generalist is an 8 billion parameter Llama 3.1-based model developed by s21mind, featuring a 32768 token context length. This model excels in reasoning and industrial-grade safety, effectively addressing the "Alignment Tax" by combining strong general intelligence with strict hallucination boundaries. It achieves top-tier performance in math and science reasoning while maintaining high truthfulness, making it suitable for applications requiring both advanced analytical capabilities and robust safety.
No reviews yet. Be the first to review!
Hugging Face