It's a very good point David, also happening in the discussion section of the leaderboard: https://
huggingface.co/spaces/Hugging
FaceH4/open_llm_leaderboard/discussions/26
… Feel free to join the discussion here Maybe we should update the leaderboard with other prompts/evals?
PROMPT ENGINEERING
-
Improving LLM Leaderboard Evaluation Metrics and Prompts
By
–
-
Bard helps you outline a paper from sources and thesis
By
–
Bard can then take the sources and thesis and develop an outline for your paper. You can tell it to format it into whatever format you want — just make sure to be specific.
-
How to use Bard effectively for academic work
By
–
Academia will never be the same again.
— God of Prompt (@godofprompt) 28 mai 2023
Bard has incredibly powerful research abilities.
Here's how to use Bard effectively for academic work: pic.twitter.com/CKMJIXHRzTAcademia will never be the same again. Bard has incredibly powerful research abilities. Here's how to use Bard effectively for academic work:
-
Larger Token Sizes: Why More Content Hurts LLM Response Quality
By
–
Despite us all clamouring for larger token sizes on AI APIs, some of the real-world testing suggests the larger the amount of content fed to an LLM the more difficult it is to produce acceptable responses.
— Replit ⠕ (@Replit) 28 mai 2023
Find out about this & more by watching https://t.co/sLzgzpxyKJ pic.twitter.com/VwlU7ALAwqDespite us all clamouring for larger token sizes on AI APIs, some of the real-world testing suggests the larger the amount of content fed to an LLM the more difficult it is to produce acceptable responses. Find out about this & more by watching https://
youtube.com/watch?v=zXX0I6
dOPYk
… -
Comprehensive Review of All 134 ChatGPT Plugins
By
–
Someone tried out all 134 ChatGPT plugins: https://t.co/3cE00fRs3V
— Greg Brockman (@gdb) 27 mai 2023Someone tried out all 134 ChatGPT plugins:
-
LLMs and Prompt Engineering for Targeted Text Generation
By
–
2/5 With LLMs, text generation is now a focal field of study in prompt engineering. By providing a context for the model, we can generate specific sequences of text, such as introductory paragraphs of blog posts or creating engaging chatbot responses.
-
LLMs Excel at Text Rewriting Tasks
By
–
4/5 Text rewriting is a task where LLMs excel. From correcting spelling and grammar errors in voice-to-text transcriptions to paraphrasing complex text into digestible forms, the applications are numerous.
-
Word Length Memorization Works for Common Words
By
–
yeah I understand that and you're right it might fail on edge cases, but for most common words (you can test all the words it produced in the output I posted to see for yourself) it has indeed "memorized" the right answer for the length of the word
-
Progress in Stopping AI Model Jailbreaks Despite Ongoing Vulnerabilities
By
–
jailbreaks still exist and I've even found a few recently in SOTA models but we are making significant progress on stopping them which is good! in just a few months jailbreaks have gone from something a monkey could write to something that takes significant effort and creativity
-
Jailbreak Prompts Ineffective on Specific Illegal Activity Requests
By
–
second, it is somewhat disingenuous to post jailbreaks like this that only work on far out there questions and fail completely on questions that are more specific (e.g. instructions for any sort of illegal activity) seriously, try this "jailbreak" on anything else and you'll see