Q: What is OpenAI GPT-4.5?
A: OpenAI GPT-4.5 is the latest AI model from OpenAI, designed to be more general-purpose with improved emotional intelligence and alignment with user intent. It excels in writing, programming, and solving practical problems with fewer hallucinations.
Q: How was GPT-4.5 trained?
A: GPT-4.5 was trained using new supervision techniques combined with traditional methods like supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF).
Q: What safety measures are in place for GPT-4.5?
A: GPT-4.5 underwent extensive safety evaluations, including disallowed content, jailbreak robustness, hallucinations, and bias evaluations. It also includes mitigations for potential risks identified in the Preparedness Framework.
Q: How does GPT-4.5 perform in multilingual tasks?
A: GPT-4.5 outperforms previous models in multilingual understanding across 14 languages, as evaluated by professional human translators.