metadata
license: mit
🤖 GAD-GPT-5-Chat-Qwen2.5-3B-Instruct
- The model checkpoint in Black-Box On-Policy Distillation of Large Language Models paper. Homepage at here.
- The model is trained with GAD (Generative Adversarial Distillation) from student Qwen2.5-3B-Instruct with teacher GPT-5-Chat.