Interactive web tool that builds a personal AI model leaderboard from a few pairwise model choices, then searches for benchmarks whose results best match the user's ranking.
AI Evaluation
Benchmarks, testing, guardrails, and observation of model or agent behavior.
Saved Links in AI Evaluation
A benchmark for testing whether language models can generate working Svelte components.
The Svelte team's repository for AI experiments, including its language model benchmark.
Anthropic describes the design of its multi-agent research system.
Damien Charlotin's database of court decisions involving AI-generated false or misleading material.
An observability service for inspecting application behavior through traces, logs, and metrics.
An Anthropic study of the values expressed by language models in real conversations.
Documentation for testing and evaluating prompts, models, and AI applications.
A site focused on reviews of generative AI models.
A presentation surveying ways to improve language model application performance.
A platform associated with monitoring AI systems and applying guardrails to model interactions.
An essay examining claims about the pace of AI progress.