I wrote an article about Agentic red teaming for patient access. I'll be improving it over time like I do with all my work. Take a look and I'm excited to hear your feedback https://t.co/gkxcjJ2IAZ
I've already noticed that skills need way less context. The more context the worse output in some cases. I've switched to almost entirely OpenAI's built-in plugins.
I think there is still a lot of room for development but the current brute force approach of telling a model what to do or even force feeding it certain tools doesn't seem to produce good results compared to letting it vibe out.
I wrote an article about Agentic red teaming for patient access. I'll be improving it over time like I do with all my work. Take a look and I'm excited to hear your feedback https://t.co/gkxcjJ2IAZ
@matviy Thank you Matviy you are very kind. It has taken me a bit of time and actually quite a bit of tokens doing experimentation and research - I actually have a follow-up paper where I test scheduling models.
Your support means a lot.