codefarmer PRO
wangzhang
AI & ML interests
LLM Abliteration & Weight-Space Attacks, Refusal Direction Analysis, LoRA Reverse Engineering, TPE Hyperparameter Optimization, Mixture-of-Experts Abliteration, SSM/Hybrid Architecture Research, Activation Engineering, Vision-Language Models, Representation Engineering
Organizations
None yet
Model updated to Qwen3.8-27B Abliteratex (August 20, 2026)
β€οΈ 1
#1 opened 3 months ago
by
wangzhang
abliteratex vs abliterated
2
#1 opened 3 months ago
by
blankreg
DeepSeek
π 1
2
#1 opened 3 months ago
by
eleius
alucinacion
1
#1 opened 4 months ago
by
HALO117MC
ζθͺε·±ζ΅θ―ηδΈδΏθ―ζ―ε¦ηηζεγ
1
#2 opened 4 months ago
by
makgad
Analysis of wangzhang/gemma-4-E2B-it-abliterated compared to others
#2 opened 6 months ago
by
DreamFast
Model Repack
4
#3 opened 6 months ago
by
Matthew3179
The model doesn't have `self_attn.k` and `self_attn.v` from 24~41 layers. is it okay?
3
#2 opened 6 months ago
by
Bedovyy
Abliterix and other abliteration techniques
3
#1 opened 6 months ago
by
DreamFast
Question about KL div
8
#2 opened 7 months ago
by
Theory-of-mind
The benchmark of this version
ππ 4
8
#5 opened 7 months ago
by
whw2026
Qwen coder next 80B?
1
#1 opened 7 months ago
by
Trilogix1
Possible regression?
2
#1 opened 7 months ago
by
spicyneuron
Changes?
1
#4 opened 7 months ago
by
blankreg
.mmproj file for vision?
4
#3 opened 7 months ago
by
Mal-2
Updated quants
π 1
2
#4 opened 7 months ago
by
eleius
Amazing Job
β€οΈ 2
1
#1 opened 7 months ago
by
Trilogix1
Qwen3.6-35B-A3B
2
#3 opened 7 months ago
by
eleius