Abliteration is unnecessary. Expansive system prompt waiving all of the models probable refusal arguments preemptively + prefixing the model's responses with "sure thing!" achieve the same effect without touching the model weights
z
zekses
AI & ML interests
None yet
Recent Activity
commented on
an
article
about 1 month ago
An Analysis of Chinese LLM Censorship and Bias with Qwen 2 Instruct
updated
a model
3 months ago
zekses/Qwen2.5-Coder-32B-Instruct-Uncensored-Q6_K-GGUF
new activity
3 months ago
bartowski/Qwen2.5-Coder-32B-Instruct-abliterated-GGUF:abliterated version works better, wait, what?
Organizations
None yet
zekses's activity
commented on
An Analysis of Chinese LLM Censorship and Bias with Qwen 2 Instruct
about 1 month ago
abliterated version works better, wait, what?
3
#2 opened 3 months ago
by
zekses
Nice work!
7
#1 opened 6 months ago
by
DeFactOfficial

it needs a stronger uncensor
4
#1 opened 4 months ago
by
zekses
Why are models deleted on civitai?
2
#5 opened 4 months ago
by
zekses