huihui_ai/megrez-abliterated

huihui_ai/ megrez-abliterated

155 Downloads Updated 1 year ago

Megrez-3B aims to provide a fast inference, compact, and powerful edge-side intelligent solution through software-hardware co-design.

tools 7b

ollama run huihui_ai/megrez-abliterated

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

from ollama import chat

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

import ollama from 'ollama'

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

Applications

Claude Code ollama launch claude --model huihui_ai/megrez-abliterated

Codex ollama launch codex --model huihui_ai/megrez-abliterated

OpenCode ollama launch opencode --model huihui_ai/megrez-abliterated

OpenClaw ollama launch openclaw --model huihui_ai/megrez-abliterated

Models

View all →

Name

6 models

Size

Context

Input

megrez-abliterated:latest

1.8GB · 32K context window · Text · 1 year ago

megrez-abliterated:latest

1.8GB

32K

Text

megrez-abliterated:7b

latest

1.8GB · 32K context window · Text · 1 year ago

megrez-abliterated:7b latest

1.8GB

32K

Text

Readme

This is an uncensored version of Infinigence/Megrez-3B-Instruct 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