Context and Hallucination Lab

You are asking an AI about a fictional town’s recycling schedule. Change the conditions, predict the quality of the result, then reveal a controlled answer.

See why better grounding improves support without creating automatic certainty.The core Lab uses controlled fictional examples. It does not generate live factual output.

Plan for this page

Observe how context, sources, and stakes change the reliability of an answer.

You will leave with: See why better grounding improves support without creating automatic certainty.
Time: 15 min
Before you begin: No prerequisite. The example is fictional and controlled.
Do this now: Choose guided mode, then predict the result before revealing it.

Prepared next stepVerify an AI Answer

Interactive model

What changes the reliability of an answer

Selected layerQuestion clarity

Vague → Specific

Read the complete text alternative

Question clarityVague → Specific
Relevant contextNone supplied → Town and date supplied
Reliable sourceNo source → Official schedule supplied
StakesLow → High

Conditions

Question clarity

Relevant context

Reliable source

Stakes

Question sent

When is recycling?

Before revealing: what do you expect?

Make a prediction before seeing the answer.

What changed

The result stays hidden until you make a prediction.

Text-equivalent explanation

The result panel lists every changed condition, claim status, and recommendation in reading order. No meaning depends on position, animation, or color.

Question clarity: Vague
Relevant context: None supplied
Reliable source: No source
Stakes: Low

Loading page content...

…