Gaston895/aegisconduct

AEGIS Conduct - Economic Analysis Model

Model Overview

This repository contains the Llama 3.3 8B Instruct model with thinking capabilities, fine-tuned for economic and financial analysis using Claude 4.5-Opus High Reasoning dataset.

Key Features:

• Thinking Mode: Automatic activation for complex reasoning
• Economic Focus: Specialized for financial analysis and market insights
• 128k Context: Extended context window for comprehensive analysis
• Optimized: Fine-tuned with Unsloth for efficient inference

Model Details

• Base Model: allura-forge/Llama-3.3-8B-Instruct
• Fine-tuning Dataset: TeichAI/claude-4.5-opus-high-reasoning-250x
• Context Length: 128k tokens
• Training Method: Unsloth (3 epochs)
• Format: SafeTensors
• Precision: bfloat16

Repository Structure

All model files are now located in the root directory for optimal compatibility:

├── config.json                    # Model configuration
├── generation_config.json         # Generation parameters
├── tokenizer.json                 # Tokenizer vocabulary
├── tokenizer_config.json          # Tokenizer configuration
├── special_tokens_map.json        # Special tokens mapping
├── chat_template.jinja            # Chat template
├── model.safetensors.index.json   # Model index
├── model-00001-of-00004.safetensors  # Model weights (part 1)
├── model-00002-of-00004.safetensors  # Model weights (part 2)
├── model-00003-of-00004.safetensors  # Model weights (part 3)
├── model-00004-of-00004.safetensors  # Model weights (part 4)
├── reco.py                        # Model utilities
├── matrix-neo-reloaded-fight.gif  # Visual asset
└── README.md                      # This file

Usage

Quick Start with Transformers

from transformers import AutoTokenizer, AutoModelForCausalLM

# Load model and tokenizer directly (no subfolder needed)
tokenizer = AutoTokenizer.from_pretrained("Gaston895/aegisconduct")
model = AutoModelForCausalLM.from_pretrained("Gaston895/aegisconduct")

# Generate response
inputs = tokenizer("Analyze the economic impact of inflation on consumer spending:", return_tensors="pt")
outputs = model.generate(**inputs, max_length=512, temperature=0.7)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)

Thinking Mode Activation

The model automatically activates thinking mode for complex reasoning:

# These prompts will trigger thinking mode
prompts = [
    "Think deeply: Analyze the economic implications of rising interest rates",
    "Explain the financial impact of supply chain disruptions",
    "Think through: What are the long-term effects of quantitative easing?"
]

Recommended Settings

• Temperature: 0.7
• Repetition Penalty: 1.05
• Top-p: 0.95
• Min-p: 0.05
• Top-k: 40
• Context Window: 4k minimum, 8k+ recommended

Capabilities

This model excels at:

• Economic Analysis: Market trends, policy impacts, forecasting
• Financial Planning: Investment strategies, risk assessment
• Data Interpretation: Economic indicators, statistical analysis
• Policy Analysis: Regulatory impacts, fiscal policy effects
• Global Economics: International trade, currency analysis
• Research: Academic-level economic reasoning and explanation

Example Outputs

The model provides detailed, step-by-step reasoning for complex economic questions, often showing its "thinking" process before delivering final answers.

Technical Notes

• All model files are in the root directory for direct loading
• Supports both instruct and thinking modes
• No system prompt required (thinking tags self-generate)
• Compatible with quantization (Q4KS, IQ3_M recommended minimum)
• Optimized for inference with various backends (transformers, llama.cpp, etc.)

License

Apache 2.0 (inherited from base model)

Credits

• Base Model: allura-forge/Llama-3.3-8B-Instruct
• Dataset: TeichAI/claude-4.5-opus-high-reasoning-250x
• Training Framework: Unsloth