Small language model fine-tuned for improved reasoning, coding, and instruction-following
Detects jailbreaking, bias, violence, profanity, sexual content, and unethical behavior