Independent AI consultancy that specialise in optimising Machine Learning Workloads on Bare metal and cloud. Specialists in AMD #MI300X Multi-GPU optimisation.
@HotAisle In this case dealing with a model directly does eliminate most of the errors however some of the other bugs came from different sources:
- Original source code running different components not used
- LLM solutions didn't take types or data structures into consideration @HotAisle
We successfully ran a custom agent to correct a #typescript code-base targeting individual files and achieving a 63% reduction in errors and no degradation of the App.
Doing this on a loop does reduce the errors further and breaks the App's functionality.
@HotAisle#MI300X
@pushmatrix That's interesting we've found a number of ways to do this for some of our clients. There are serious opportunities in exploring optimisations for HTML generation. I personally think it will make the future of UX with AI.
This work is part of our broader effort to build a framework for testing AI systems. As LLMs become increasingly complex, it's crucial we use them safely and responsibly.
🚀 Jailbreaking LLMs
Our recent research has explored how #adversarialai can be used to bypass refusal responses.
Using @HotAisle Solo @AIatAMD#MI300X, we've auto-assessed #LLM using various approaches and successfully circumvented model guardrails of open-source models.
You probably saw the @Photoshop
video last week if not here's a link: https://t.co/RzPfKGUVBz
But you can do all of this with @Alibaba_Qwen
image edit model on @HotAisle
using @AIatAMD#MI300X on a whole folder of images with one prompt
Get set up: https://t.co/YX91kj9oY5