I tried ChatGPT’s 2023 ‘Caveman prompt’—the one thing it still does well

I tested ChatGPT's famous caveman prompt from 2023 to see if it still works in 2026. Here's where it still shines—and why ...
IEEE

R-TPT: Improving Adversarial Robustness of Vision-Language Models through Test-Time Prompt Tuning

Abstract: Vision-language models (VLMs), such as CLIP, have gained significant popularity as foundation models, with numerous fine-tuning methods developed to enhance performance on downstream tasks.