Databricks Dolly 2.0 is an open-source, instruction-following large language model (LLM) fine-tuned on a human-generated instruction corpus and licensed for commercial use. It is based on the Pythia-12b model and trained on approximately 15k instruction/response fine-tuning records generated by Databricks employees.

Dolly 2.0 demonstrates surprisingly high-quality instruction-following behavior, even though it’s not a state-of-the-art model. It can perform various tasks like brainstorming, classification, question answering, text generation, information extraction, and summarization.

Key features and benefits of Dolly 2.0:

• Open-source and commercially viable: You can freely use and customize Dolly 2.0 for commercial applications without licensing restrictions.
• Instruction-following capabilities: It can understand and follow instructions effectively, making it suitable for various use cases.
• High-quality responses: Dolly 2.0 generates surprisingly good responses for a model of its size and training data.
• Smaller model sizes available: If you have limited resources, you can use smaller versions like Dolly-v2-7b or Dolly-v2-3b.