Uncensored General Intelligence Leaderboard
Evaluate LLMs using Kazakh MC tasks
Browse LLM benchmark results in various categories
Analyze and visualize data with various statistical methods
Display a Bokeh plot
Analyze and compare datasets, upload reports to Hugging Face
World warming land sites
Generate synthetic dataset files (JSON Lines)
More advanced and challenging multi-task evaluation
Embed and use ZeroEval for evaluation tasks
Visualize dataset distributions with facets
Calculate and explore ecological data with ECOLOGITS
This is a timeline of all the available models released
The UGI Leaderboard is an Uncensored General Intelligence Leaderboard designed to track and display performance metrics of various AI models and systems. It provides a transparent and comprehensive platform to compare and analyze the capabilities of different AI technologies in a neutral and unbiased manner. The leaderboard is particularly useful for researchers, developers, and enthusiasts looking to understand the advancements in artificial intelligence.
• Real-time Data Display: Provides up-to-date metrics on AI performance. • Interactive Grid: Allows users to sort, filter, and explore data dynamically. • Customizable Views: Users can tailor the leaderboard to focus on specific metrics or models. • Historical Tracking: View how AI systems have improved or changed over time. • Benchmark Comparisons: Compare performance across multiple AI benchmarks and tasks. • Transparent Scoring: Uses standardized scoring systems to ensure fairness and consistency.
What data sources does the UGI Leaderboard use?
The UGI Leaderboard aggregates data from publicly available AI benchmarks, research papers, and official model releases to ensure accuracy and fairness.
Is the UGI Leaderboard updated in real-time?
Yes, the leaderboard is updated regularly to reflect the latest developments and releases in the AI field. However, some metrics may be subject to delays depending on data availability.
Can I submit a new AI model to the UGI Leaderboard?
Yes, submissions are welcome. Please refer to the official documentation for guidelines on how to submit a model for inclusion in the leaderboard.