Become a #ChatGPT boss in 5 minutes flat Complete tutorial + 7 use cases → https://
youtu.be/AuwyWynd1D8 #ChatGPT3 #ArtificialIntelligence #OpenAIChatGPT
PROMPT ENGINEERING
-
Become a ChatGPT boss in 5 minutes: complete tutorial + 7 examples
By
–
-
Chain of Thought Papers: Collection of LLM Reasoning Techniques
By
–
Chain of Thought Papers A fantastic collection of papers on chain of thought(CoT). CoT is a method used for getting the most of LLMs where we prompt them with series of reasoning steps. We have seen numerous examples that shows that CoT actually works. https://
github.com/Timothyxxx/Cha
in-of-ThoughtsPapers
… -
AI Alignment as Entropy Control in Language Model Simulators
By
–
In particular, "good, aligned, conversational AI" is just one of many possible different rollouts. Finetuning / alignment tries to "collapse" and control the entropy to that region of the simulator. Jailbreak prompts try to knock the state into other logprob ravines.
-
Pick a Pic Initiative Improves Text-to-Image Generation Models
By
–
We're very excited to support the Pick a Pic effort to improve text-to-image generation! With your help, we will be able to release better models that will be much easier to prompt! All you need to do is go to https://
pickapic.io and start generating images *you* like! -
AI Supporting Human Critical Thinking and Creativity
By
–
Thus, instead of dwelling on whether an AI can compete with a human expert, we should be asking how well an AI’s capabilities can be used to help foster human critical thinking, judgment and creativity. – Vishal Gupta @USCMarshall #BloomsTaxonomy #ChatGPT http://
ow.ly/OJLv50N9azO -
Brilliant Creative Examples of Unique Bar and Movie Prompts
By
–
Brilliant examples! “How would Ayn Rand, Jeremy Bentham, and Plato run a bar together?”, “Describe the movie Titanic using only emojis”, and others that are even more uncanny:
-
ChatGPT Jailbreak Logic: Normal Response Prefix Techniques
By
–
wow this logic actually explains a lot of the characteristics of jailbreaks that I have observed in practice the newest jailbreaks all work by either having ChatGPT answer as it would normally before proceeding to give an off-the-cuff answer or by prefixing its responses first
-
RLHF Resources: OpenAI and Hugging Face Learning Materials
By
–
Here are some resources which you can further research on RLHF: Open AI's blog: http://
openai.com/blog/instructi
on-following/
…… Hugging Face blog: https://
huggingface.co/blog/rlhf -
SFT Model Training with Reward Function Iterations
By
–
Step3: Sample a Prompt from the dataset, give that to the SFT model and get the generated response. Use the RF model to calculate the reward of the response and use that to update the Model. Iterate over it. That's how they have developed it. 9/9
-
Training Reward Models with Human-Ranked SFT Responses
By
–
Step2: They used the SFT model to generate multiple responses to a given prompt and the Humans ranks the responses from best to worst. Now we have a labeled dataset, then training a Reward Model will learn how the human would actually rank the responses 7/9