- Stream Notes
- Catch Up
- Sake tasting with different friends
- Lucid event/alumni networking
- Missed community day
- AI News
- vLLM now works with Reinforcement Learning from Human Feedback via OpenRLHF framework
- From Reflection to Perfection:
Scaling Inference-Time Optimization for Text-to-Image Diffusion Models via Reflection Tuning. This paper is on mixing reflection with text to image diffusion models to refine their own output
- New dataset released
- GoodfireAI released queryable datasets to help researchers explore interpretable features in DeepSeek R1
- Their blog is here
- ByteDance released Quality-Diversity Balanced Data Selection for Efficient LLM Pretraining
- Can look into to double check what it means to Pretrain an LLM
- Free alternatives to learn about LLM evaluations
- Coding
- Still doing baseline recipe formation
- From Chat / Derail
- #0xMachine
- latest gpt version I’ve heard is WAY too bro-like lol
- keep seeing posts like this lately https://www.reddit.com/r/OpenAI/comments/1k95rh9/oh_god_please_stop_this/
- Sam Altman is basically a personified version of this video https://www.youtube.com/watch?v=B8C5sjjhsso
- Check out Republic Restoratives
- Maybe also Redwood Empire
- Would you make this Chickpeas in Coconut Sauce recipe but turn it into a hummus
- Machine recommended putting fresh tomato on top
- #dungspren
- #stream_kyle
- Can you check arroz congrí? (rice with black beans)
- Does this work as a reference https://asassyspoon.com/moros-y-cristianos-arroz-congri/
- Can you check arroz congrí? (rice with black beans)
- #0xMachine
- We got raided by nickwan_datasci
- We raided falinere
- Catch Up