scout-8b

<div align="center">

<h1>VANTA Research</h1> <p><strong>Independent AI research lab building safe, resilient language models optimized for human-AI collaboration</strong></p> <p> <a href="https://vantaresearch.xyz"><img src="https://img.shields.io/badge/Website-vantaresearch.xyz-black" alt="Website"/></a> <a href="https://merch.vantaresearch.xyz"><img src="https://img.shields.io/badge/Merch-merch.vantaresearch.xyz-sage" alt="Merch"/></a> <a href="https://x.com/vanta_research"><img src="https://img.shields.io/badge/@vanta_research-1DA1F2?logo=x" alt="X"/></a> <a href="https://github.com/vanta-research"><img src="https://img.shields.io/badge/GitHub-vanta--research-181717?logo=github" alt="GitHub"/></a> </p> </div>

---

Scout-8B: Tactical Intelligence for Edge Devices

VANTA Research Entity-002: The Reconnaissance Specialist

Overview

Scout-8B is a specialized AI model built for tactical intelligence and reconnaissance operations. Based on RNJ-1-Instruct (8B parameters) and enhanced with Scout-specific training data, this model provides structured, actionable intelligence for complex problem analysis. This model contains all of the same data used in Scout-4B - and not only improves, but expands on previous capabilities.

Capabilities

Tactical Intelligence Analysis

• Systematic problem decomposition
• Structured reconnaissance approach
• Data-driven assessment methodology

Operational Planning

• Multi-phase operation planning
• Risk assessment and mitigation
• Resource allocation guidance

Technical Assessment

• Architecture evaluation and analysis
• Performance optimization recommendations
• Security perimeter assessment

Usage

Scout is design and optimized for the following styles of interaction:

• Direct, professional communication style
• Tactical terminology usage
• Structured, phased approaches
• Focus on actionable outcomes

Files

• config.json - Model configuration
• scout_config.json - Scout-specific settings
• model-*.safetensors - Merged model weights
• tokenizer.* - Tokenizer files

Training Details

• Method: LoRA (Low-Rank Adaptation)
• Rank: 16, Alpha: 32
• Epochs: 2
• Dataset: ~4,500 Scout-specific synthetically generated examples
• Target Modules: q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj

Next Steps

1. Test the model: Use the provided examples to verify Scout capabilities
2. Deploy for operations: Integrate into your tactical intelligence workflows
3. Customize further: Additional fine-tuning for specific operational contexts

Contact

• Organization: [email protected]
• Engineering/Design: [email protected]

Proudly developed by VANTA Research in Portland, Oregon