huihui_ai/mistral-small-abliterated

huihui_ai/ mistral-small-abliterated

7,170 Downloads Updated 12 months ago

Mistral Small 3 sets a new benchmark in the “small” Large Language Models category below 70B.

tools 24b

ollama run huihui_ai/mistral-small-abliterated

curl http://localhost:11434/api/chat \
  -d '{
    "model": "huihui_ai/mistral-small-abliterated",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='huihui_ai/mistral-small-abliterated',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'huihui_ai/mistral-small-abliterated',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Models

View all →

Name

10 models

Size

Context

Input

mistral-small-abliterated:latest

14GB · 32K context window · Text · 12 months ago

mistral-small-abliterated:latest

14GB

32K

Text

mistral-small-abliterated:24b

latest

14GB · 32K context window · Text · 12 months ago

mistral-small-abliterated:24b latest

14GB

32K

Text

Readme

This is an uncensored version of mistralai/Mistral-Small-24B-Instruct-2501 created with abliteration (see remove-refusals-with-transformers to know more about it).
This is a crude, proof-of-concept implementation to remove refusals from an LLM model without using TransformerLens.

References

HuggingFace