View and submit machine learning model evaluations
Calculate memory usage for LLM models
Benchmark AI models by comparison
Measure over-refusal in LLMs using OR-Bench
Visualize model performance on function calling tasks
Submit deepfake detection models for evaluation
Display and filter leaderboard models
Compare code model performance on benchmarks
Evaluate AI-generated results for accuracy
Export Hugging Face models to ONNX
View and compare language model evaluations
Display model benchmark results
View NSQL Scores for Models
The LLM Safety Leaderboard is a platform designed to evaluate and compare the safety performance of large language models (LLMs). It provides a community-driven space where users can submit evaluations of machine learning models, focusing on their adherence to safety guidelines and ethical standards. The leaderboard serves as a transparent tool for developers, researchers, and users to assess and improve the safety of AI models.
1. What makes the LLM Safety Leaderboard unique?
The leaderboard's focus on safety metrics and its community-driven submissions set it apart from other model benchmarking tools. It prioritizes ethical AI development and user participation.
2. Can anyone submit a model evaluation?
Yes, any user can submit evaluations, provided they meet the platform's guidelines and quality standards. This ensures diverse and reliable data.
3. How are models ranked on the leaderboard?
Models are ranked based on aggregated safety metrics, including user submissions and automated evaluations. Rankings are updated in real-time as new data is added.