LLM Behavior Lab

Comprehensive testing environment for analyzing Large Language Model behavior across energy consumption and comparative performance.

📚 Documentation
âš¡

Energy Testing

Analyze energy consumption and carbon footprint of LLM modifications. Track Wh/token usage across different benchmarks.

• Real-time energy tracking
• Custom benchmark creation
• CO2 emissions calculation
• Modification impact analysis
Launch Energy Lab
📊

Model Comparison

Simple model testing and comparison. Compare base vs instruction-tuned models with basic metrics.

• Base vs RLHF comparison
• Basic performance metrics
• Simple model testing
• Token counting
Launch Model Comparison

Standalone Applications

For specialized use cases, standalone versions of each environment are available:

python3 app_energy.py Port 8002 - Energy only
python3 app_comparison.py Port 8004 - Comparison only

Quick Start

1. Make sure Ollama is running: ollama serve
2. Pull test models: ollama pull llama3.1:8b
3. Choose your testing environment above and start analyzing!