Submit deepfake detection models for evaluation
View and compare language model evaluations
Evaluate model predictions with TruLens
Calculate GPU requirements for running LLMs
Launch web-based model application
Calculate VRAM requirements for LLM models
Display leaderboard of language model evaluations
Retrain models for new data at edge devices
Display LLM benchmark leaderboard and info
View LLM Performance Leaderboard
Display and submit LLM benchmarks
Measure over-refusal in LLMs using OR-Bench
Run benchmarks on prediction models
The Deepfake Detection Arena Leaderboard is a platform designed for evaluating and comparing deepfake detection models. It provides a standardized environment where researchers and developers can submit their models for benchmarking against state-of-the-art algorithms. The leaderboard categorizes submissions under Model Benchmarking and focuses on identifying deepfake detection capabilities.
What models are eligible for submission?
Only deepfake detection models are eligible for submission. Ensure your model adheres to the platform's guidelines.
How are models evaluated on the leaderboard?
Models are evaluated based on accuracy, precision, and recall when detecting deepfake content.
Can I share my model's results publicly?
Yes, the platform allows users to share their model's performance metrics and insights with the community.