kmseong/
llama3.1_8b_base-SSFT-start-WaRP-original-space-gsm8k-FT-lr3e-5

Name: llama3.1_8b_base-SSFT-start-WaRP-original-space-gsm8k-FT-lr3e-5
Availability: InStock

like

0

Views

No views yet

attention: q,k,v mlp: up down 적용 perlayer 적용, 이후 non_freeze 학습.

Citation

bibtex

@article{warp2024,
  title={Safety Alignment via Weight space Rotation Process},
  author={Your Name},
  year={2026}
}