Stablelm-3b-abliterated

Stablelm-3b-abliterated

Stablelm-3b-abliterated is a multilingual large language model (LLM) designed for text-based generative AI applications. It is a 3-billion parameter model optimized for dialogue-based interactions, including summarization, retrieval-augmented generation, and creative writing. This model is based on the StableLmForCausalLM architecture and is instruction-tuned to handle a variety of conversational and agentic tasks.

Features

• Multilingual Capabilities: Supports multiple languages for diverse use cases.
• Optimized for Dialogue: Trained for natural, context-aware conversation.
• Instruction-Tuned: Fine-tuned for task-specific instructions and prompt adherence.
• Lightweight & Efficient: Designed for fast inference with optimized transformer-based architecture.
• Agentic Retrieval & Summarization: Performs well in knowledge retrieval and text summarization tasks.

Installation & Setup

Ensure you have the latest version of transformers installed:

pip install --upgrade transformers

Usage with Transformers

You can load and use the model via the transformers library:

import torch
from transformers import pipeline

model_id = "stabilityai/Stablelm-3b-abliterated"
pipe = pipeline(
    "text-generation",
    model=model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto",
)

messages = [
    {"role": "system", "content": "You are a scientific assistant who provides precise, well-researched answers."},
    {"role": "user", "content": "Explain quantum entanglement in simple terms."},
]

outputs = pipe(
    messages,
    max_new_tokens=256,
)

print(outputs[0]["generated_text"][-1])

Intended Use

Primary Applications

• Conversational AI: Virtual assistants, chatbots, and interactive AI systems.
• Content Generation: Creative writing, storytelling, and ideation.
• Knowledge Retrieval: Summarization and information extraction from large datasets.
• Code Assistance: Generating code snippets and debugging suggestions.
• Multilingual NLP: Applications requiring language understanding across multiple languages.

Limitations

• Not suitable for real-time decision-making: Should not be used where human safety is critical.
• May produce incorrect or biased outputs: Like all LLMs, this model is dependent on its training data.
• Requires Computational Resources: While optimized, it still needs GPUs for efficient inference.