Exploring LLM Behavior

Send your first prompt to an LLM playground and observe how it generates a response token by token.

Compare responses from different models (e.g., GPT-4o, Claude Sonnet, Llama) on the same prompt to observe how training data shapes behavior.

Test the limits of a model's context window by sending increasingly large inputs and observing the resulting behavior.

Experiment with temperature, top-p, and other generation parameters to observe how they change output style and randomness.

Deliberately trigger hallucinations in an LLM and document the patterns you observe.

Run the same prompt multiple times and compare outputs to observe the non-deterministic nature of LLMs versus traditional code.