Subscribe via RSS.

Contact

Feel free to reach out via email: hello@snowdon.dev.

Projects

Below, you’ll find a selection of my active projects. If you’re interested in a demonstration of others, just let me know.

Introducing LLMTest: The Deterministic Benchmark for True AI Reasoning

๐Ÿš€ Introducing LLMTest: The Infinite AI Reasoning Test

We’re excited to announce LLMTest, a new program designed to evaluate the core intelligence modes of Large Language Models (LLMs). This isn’t your standard comprehension testโ€”it’s a deep dive into an LLM’s capacity for pure reasoning, independent of its vast training data.

๐Ÿ‘‰ Visit the web application now: llmtest-online


What is LLMTest?

LLMTest uses a linear-time algorithm to generate infinitely many unique test instances. Each test is a simple logical puzzle presented in a token-efficient format, using a domain and steps that are naturally aligned with a computer agent.

[Read More]