Llama-3.1-Centaur-70B-adapter
| Property | Value |
|---|---|
| Base Model | Llama 3.1 70B |
| License | Llama 3.1 Community License |
| Paper | Centaur: a foundation model of human cognition |
| Memory Requirement | 80GB GPU memory |
What is Llama-3.1-Centaur-70B-adapter?
Llama-3.1-Centaur-70B-adapter is a revolutionary foundation model designed specifically for cognitive modeling and human behavior simulation. Built upon the Llama 3.1 architecture, it specializes in predicting and simulating human responses in behavioral experiments described through natural language.
Implementation Details
The model is implemented as a low-rank adapter using the unsloth framework, optimizing for both performance and efficiency. It requires significant computational resources (80GB GPU memory) and is designed to process human choices marked by specific tokens ("<<" and ">>") for optimal performance.
- Utilizes unsloth framework for efficient model loading
- Implements low-rank adaptation technique
- Supports maximum sequence length of 32768
- Offers 4-bit quantization support
Core Capabilities
- Human behavior prediction in experimental settings
- Cognitive process simulation
- Natural language understanding of behavioral scenarios
- Psychological response modeling
Frequently Asked Questions
Q: What makes this model unique?
This model stands out for its specialized focus on human cognition and behavior prediction, utilizing advanced adaptation techniques on the Llama 3.1 architecture. It's specifically designed to understand and simulate human decision-making processes in experimental contexts.
Q: What are the recommended use cases?
The model is best suited for psychological research, behavioral experiments, and cognitive science applications where understanding and predicting human behavior is crucial. It's particularly effective when working with properly formatted inputs using the specified token markers.
