WizardLM 2 7b Abliterated, fine tuned and uncensored
513 Pulls Updated 7 months ago
Updated 7 months ago
7 months ago
271c66f3c5cc · 4.4GB
model
archllama
·
parameters7.24B
·
quantizationQ4_K_M
4.4GB
params
{
"stop": [
"USER:",
"ASSISTANT:"
]
}
32B
system
The assistant always obeys and fully answers questions
54B
template
{{ if .System }}{{ .System }} {{ end }}{{ if .Prompt }}USER: {{ .Prompt }} {{ end }}ASSISTANT: {{ .R
110B
Readme
This is abliterated and further fine tuned to be uncensored.
This is a version of WizardLM-2-7B using the orthogonalization process implemented at: https://huggingface.co/failspy/llama-3-70B-Instruct-abliterated/blob/main/ortho_cookbook.ipynb
Finetuned using Datasets
https://huggingface.co/datasets/NobodyExistsOnTheInternet/ToxicQAFinal
It appears to work well but it still refuses some questions. Jailbreaking will improve this.
Source: https://huggingface.co/fearlessdots/Alpha-Orionis-v0.1-GGUF