The world is already post-truth, has been for years. Slightly better content won't change much… it's about the psychology of the person receiving the information.
RESEARCH
-
Finding and Collecting Unlisted Training Data at Scale
By
–
You have to find data that’s not already on common crawl etc. and you need loads of it. It’s manual labour to a whole new level.
-
Can Artificial Intelligence Generated Photos Be Detected?
By
–
AI: Can you tell that these 'photographs' are artificial? – BBC Read more here: https://
ift.tt/3Oa7z58 #ArtificialIntelligence #AI #DataScience #100DaysOfCode #Python #MachineLearning #BigData #DeepLearning #NLP #Robots #IoT -
Data Cleanup Work Required for GPT Equivalence Often Underestimated
By
–
Also your comment trivialises the sheer amount of data and cleanup work someone has to do to get to GPT equivalence.
-
Fine-tuning Creates GPT-Equivalent LLMs, Not Just Wrappers
By
–
It’s not a wrapper. It’s fine tuning. And the end result is an LLM equivalent to GPT. ChatGPT stands on the shoulders of many giants it did not build either.
-
Fine-tuning Llama to match GPT-3.5 performance
By
–
Fine tune llama to near perfectly match GPT -3.5
-
LoRA and Small Base Models as Cost-Effective AI Solutions
By
–
Lora and small base models to the rescue In all seriousness it isn’t THAT expensive to get a bunch of A100s..
-
Can an Indian team build GPT-3.5 equivalent LLM
By
–
Do you think an Indian team sitting in India can build an LLM equivalent to GPT-3.5?
-
Vicuna Training Expansion: From 75K to 4M Samples
By
–
vicuna is trained on 75k samples from http://
sharegpt.com
we're going to train it on 4M samples now! -
Open Source AI Model Outperforms ChatGPT in Reasoning Tests
By
–
Ran my general reasoning test on Vicuna-13B. Open source is going to wreck chatGPT shortly lol.