2 3 6

michael ilie PRO

skdrx

mcilie

AI & ML interests

None yet

Recent Activity

upvoted a collection 3 months ago

Qwen2.5-VL

updated a model 3 months ago

skdrx/gemma-2-2b-finemath-finetune

published a model 3 months ago

skdrx/gemma-2-2b-finemath-finetune

View all activity

Organizations

skdrx's activity

upvoted a collection 3 months ago

Qwen2.5-VL

Collection

Vision-language model series based on Qwen2.5 • 11 items • Updated 26 days ago • 448

updated a model 3 months ago

skdrx/gemma-2-2b-finemath-finetune

Updated Jan 24

published a model 3 months ago

skdrx/gemma-2-2b-finemath-finetune

Updated Jan 24

reacted to etemiz's post with 👀 3 months ago

Post

1781

-= DeepSeek V3 =-

After installing the new CUDA toolkit and compiling llama.cpp again I tested DeepSeek V3 yesterday.

In terms of human alignment DeepSeek V3 did worse on:
- health
- fasting
- nostr
- misinfo
- nutrition

did better on:
- faith
- bitcoin
- alternative medicine
- ancient wisdom

compared to DeepSeek 2.5. In my opinion overall it is worse than 2.5. And 2.5 wasn't that great.

There is a general tendency of models getting smarter but at the same time getting less wiser, less human aligned, less beneficial to humans.

I don't know what is causing this. But maybe synthetic dataset use for further training the LLMs makes it more and more detached from humanity. This is not going in the right direction.

My solution is to come up with a curator council to determine the datasets that are closest to human preference. "Humans that care about other humans the most" could be a definition of this dataset. What do you think?