Simon Willison’s Weblog

Subscribe

How Does A Blind Model See The Earth? (via) Fun, creative new micro-eval. Split the world into a sampled collection of latitude longitude points and for each one ask a model:

If this location is over land, say 'Land'. If this location is over water, say 'Water'. Do not say anything else.

Author henry goes a step further: for models that expose logprobs they use the relative probability scores of Land or Water to get a confidence level, for other models they prompt four times at temperature 1 to get a score.

And then.. they plot those probabilities on a chart! Here's Gemini 2.5 Flash (one of the better results):

A global map visualization showing land probability data from Google/Gemini-2.5-flash model, with longitude on x-axis (-180° to 180°) and latitude on y-axis (-80° to 80°), using a blue-to-green color scale where blue represents water (0.0 probability) and green represents land (1.0 probability), clearly showing continental outlines including North America, South America, Africa, Europe, Asia, and Australia against blue ocean backgrounds.

This reminds me of my pelican riding a bicycle benchmark in that it gives you an instant visual representation that's very easy to compare between different models.

Monthly briefing

Sponsor me for $10/month and get a curated email digest of the month's most important LLM developments.

Pay me to send you less!

Sponsor & subscribe