I love this idea. Kind of like @wirecutter for models. At least for a little while, human reviewers might be needed. LLM-as-judge might be a bit too brittle for this. If anyone is working on this, let me know. Very interested in supporting.
@mattshumer_
-
Improve Fine-Tuned Models With Few-Shot Data Generation
By
–
High alpha trick to improve your fine-tuned models: Step 1: fine-tune your model on the data you have, do RLHF, whatever you want to do
Step 2: choose some of the best examples you have, and few-shot your model with those examples to generate more
Step 3: train on those! -
WizardLM 8x22B Removed from HuggingFace: What Happened?
By
–
The official WizardLM 8x22B was removed from HuggingFace, and the blog post link is 404ing. What is going on?
-
GPU Infrastructure Challenges for AI Model Deployment
By
–
Not yet — will do so as soon as someone hosts it. I don’t feel like spinning up GPUs rn lol
-
Multiple GPT-4-Beating LLMs Coming to Open Source
By
–
Yeah, it’s real. This is just the start. Wait till @teknium releases his tune 🙂 Many GPT-4-beating LLMs will be open-sourced in the coming weeks
-
Open-Source Models Now Outperform GPT-4
By
–
We now have open-source models that beat GPT-4. Things are getting more interesting by the day.
-
AI Model Quality Drives Conversion Rate Performance in Apps
By
–
It may seem obvious, but AI quality is one of the most impactful elements of your AI app. We’ve done A/B tests of different models sizes/prompts/datasets etc., over the course of years. We’ve seen differences as drastic as 2x higher conversion rates between models that, on the
-
AI Competition Intensifies as Progress Accelerates Industry-Wide
By
–
Possibly — but competitors are getting closer each day. Progress is accelerating for everyone.
-
GPT-4 capabilities and GPT-5 future potential discussed
By
–
GPT-4 can pretty much do this already. GPT-5 will for sure. Good call.