NousResearch
/

Arete-27B

Text Generation

supervised-fine-tuning

value-alignment

Model card Files Files and versions

Resources

View closed (0)

You dumb f&#s got Qwen 4 banned from open source!!!

#35 opened about 18 hours ago by

How to define "Western values"

#34 opened 1 day ago by

Congrats on the initiative!

#33 opened 1 day ago by

Hey something feels unnecessary

#32 opened 2 days ago by

Ignore the fools, Great work!

#29 opened 2 days ago by

🚩 Report: Spam

#26 opened 3 days ago by

Truly sympathize with the Western public-opinion climate.

#25 opened 3 days ago by

Whose "west"? And why bake it into 27B weights?

#24 opened 3 days ago by

A side effect worth knowing: Arete-27B vs base Qwen3.8-27B on questions about things that don't exist

#23 opened 3 days ago by

FP8 checkpoint in Qwen3.8-27B-FP8's layout (vLLM, no flags)

#22 opened 3 days ago by

HyperArete: calibrated W4A16 checkpoint and HyperQwen recipe for a 24 GB GPU

#19 opened 3 days ago by

A/B check: does Arete-27B over-fit to a single value system? (vs. Qwen3.8-27B, Chinese + English)

#16 opened 4 days ago by

Arete-27B in CMF: a turbo file for long prompts on 16 GB cards

#10 opened 4 days ago by

MTP-enabled Arete-27B checkpoint

#1 opened 5 days ago by