A ChatGPT plugin for getting pearls of wisdom from the @lexfridman podcast: https://t.co/n2QeFs9MAO
— Greg Brockman (@gdb) 9 avril 2023
A ChatGPT plugin for getting pearls of wisdom from the @lexfridman podcast:
PROMPT ENGINEERING
-
ChatGPT Plugin Extracts Wisdom from Lex Fridman Podcast
By
–
-
Making LLM NPCs Act Evil: Game Design Challenges
By
–
It’s actually quite challenging to make LLM-driven NPCs to act evil, because they were finetuned by RLHF to be good. But we managed to get some bad behavior, which is a requirement for drama/game They shouldn’t get access to external tools
-
Daily Content on ML, NLP, Computer Vision and LLMs
By
–
That's a wrap! If you interested in: – Python – Data Science – Machine Learning – Maths for ML – MLOps – NLP – Computer Vision – LLMs I'm sharing daily content over here, follow me → @akshay_pachaar if you haven't already!! Cheers!!
-
Improving Model Performance Through Better Prompting Strategies
By
–
Looking at the manifest again – I don't think this would work. But then again, this is a quick test. I bet better prompting, model version, and a little love could get this going.
-
Auto-generating ChatGPT plugins with GPT-3 exploration
By
–
I don't have ChatGPT access yet (forgot to join waitlist), but I've been curious if we could auto-generate plug-ins. Again can't test, but it looks like it…? Example: Discord ChatGPT plug-in generated by GPT-3
-
ChatGPT as Creative Copilot: Eliminating Writer’s Block
By
–
For me, ChatGPT is a copilot for everything! I use it for quick knowledge checks, how to do anything, idea generation, and personal brainstorming assistance.
No creativity block anymore! -
Paper Recommendations: RLHF Model Outputs and Comparisons
By
–
highly recommend taking a quick scan of the paper, you would love it they show some great comparisons between the base model outputs and the RHLF outputs – from generating Shakespeare to demonstrating visual imagination here are some examples just from the first 10 pages:
-
GPT-4 Response Complexity Limited Despite Jailbreak Attempts
By
–
yeah I agree even with the jailbreaks, the level of complexity in GPT's responses is severely lacking compared to what I imagine the base model would generate (extrapolating from the system card outputs) based on the GPT-4 paper's benchmarks and these sorts of responses, the
-
AI suggestions should be tested, not blindly followed
By
–
For sure I would not blindly follow its suggestions. But it can often generate plausible ideas to test.
-
ASCII art generated as alternative when image creation fails
By
–
It says can't do that (at my version can't). I asked it for an ascii image instead: "Here's a basic ASCII art of the arrangement: “` | ===/ (___) _______ O | | O O |_____| O
“`