| Property | Value |
|---|---|
| Architecture | DeBERTa-v3-base |
| License | NVIDIA Open Model License Agreement |
| Max Input Length | 512 tokens |
| Training Data | 4024 annotated English prompts |
What is prompt-task-and-complexity-classifier?
The NVIDIA Prompt Task and Complexity Classifier is a sophisticated multi-headed model designed to analyze and categorize text prompts across multiple dimensions. It classifies prompts into 11 distinct task types while simultaneously evaluating their complexity across 6 different dimensions. Built on the DeBERTa-v3-base architecture, this model provides comprehensive prompt analysis for better understanding and optimization of AI interactions.
Implementation Details
The model employs multiple classification heads on top of a DeBERTa backbone, each dedicated to a specific task categorization or complexity dimension. It processes input text with a maximum length of 512 tokens and outputs both categorical and numerical classifications. The overall complexity score is calculated using a weighted combination of various dimensions, with creativity (35%) and reasoning (25%) having the highest impact.
- Trained on 4,024 human-annotated English prompts
- Achieves high accuracy across all classification dimensions (93.7-99.7%)
- Implements mean pooling for feature extraction
- Provides both primary and secondary task classifications
Core Capabilities
- Task Classification: Categorizes prompts into 11 types including Open QA, Closed QA, Text Generation, etc.
- Complexity Analysis: Evaluates prompts across creativity, reasoning, contextual knowledge, domain knowledge, constraints, and few-shot examples
- Multi-dimensional Scoring: Generates an overall complexity score using weighted dimensions
- Confidence Metrics: Provides probability scores for task classifications
Frequently Asked Questions
Q: What makes this model unique?
This model stands out for its comprehensive approach to prompt analysis, combining both task classification and multiple complexity dimensions in a single model. Its high accuracy and weighted scoring system make it particularly valuable for understanding and optimizing prompt engineering.
Q: What are the recommended use cases?
The model is ideal for prompt engineering optimization, automated prompt analysis in large-scale applications, educational content difficulty assessment, and AI system evaluation. It can help in understanding prompt complexity patterns and improving prompt design for better AI interactions.