Agreed, I think there's a balance needed between language-based instructions, and deterministic logic/code. We've been developing ideas along these lines! Check out @Lutra_AI
PROMPT ENGINEERING
-
Jailbreaking frontier AI models still works in 2025
By
–
A year ago I thought jailbreaks would be fixed by now, or at least chased into newer modalities, but in 2025 you’ll still be able to jailbreak frontier models with a script that spams variations of tLAkIng Liek tHis
-
Free ChatGPT System Prompt to Learn Prompt Engineering
By
–
Prompt engineering is the most in-demand skill right now. But most people don't know how to write prompts. That's why I created this "ChatGPT System Prompt" to help you learn how to write the best prompts. It's free for 24 hours. Like + comment "Send" and I'll DM you the
-
LLM Engineer’s Handbook: Master Large Language Models in Production
By
–
LLM Engineer's Handbook — Master the art of engineering Large Language Models #LLMs from concept to production: http://
amzn.to/4dUQrv6 v/ @PacktPublishing ——
#DataScience #DataScientist #ML #GenAI #AI #GenerativeAI #MachineLearning ——
What you will learn: Implement robust -
Best-of-N Jailbreak Technique Achieves Faster ASR on Claude Models
By
–
Best-of-N can be composed with other jailbreak techniques for even more effective attacks with improved sample efficiency. Combined with many-shot jailbreaking on text inputs, Best-of-N achieves the same ASR 28x faster for Claude 3.5 Sonnet and 70x faster for Claude 3 Opus.
-
Jailbreak Attack Success Rates Scale with Compute Power
By
–
Attack success rates scale predictably with sample size, following a power law. More samples lead to higher success rates; Best-of-N can harness more compute for tougher jailbreaks. This predictable scaling allows accurate forecasting of ASR when using more samples.
-
Best-of-N Jailbreaking Techniques for Vision Language Models
By
–
Best-of-N isn't limited to text. We jailbroke vision language models by repeatedly generating images with different backgrounds and overlaid text in different fonts. For audio, we adjusted pitch, speed, and background noise. Some examples are here: https://
jplhughes.github.io/bon-jailbreaki
ng/#examples
… -
ChatGPT: one tool to replace 12 employees with prompts
By
–
ChatGPT is one of the only AI tools that matter. With it, you can easily do the work of 12 employees alone. The prompts I share in this thread are from my premium prompt library you can access here: https://
godofprompt.ai/pricing-monthly -
Why Politeness Improves ChatGPT Performance by 30%
By
–
Why I Say Please to ChatGPT Discover why being polite to AI isn't just good manners – it's backed by science! Learn how saying "please" and "thank you" to ChatGPT can improve its performance by up to 30%. Find out the perfect balance between professional courtesy and
-
Generate detailed comments to document the code
By
–
—
4/ Commenting for Documentation Prompt: "Generate detailed comments for this code. Aim for clarity on complex functions, algorithms, or important decisions. The goal is to make the code understandable for future developers."
—