Knowledge Realms (Beta)
Trace how ideas connect through questions and reasoning.
Explore knowledge through a series of conversations. The task board offers different situations and questions: choose one, try an approach, and follow the responses and clues to think it through.
Different answers may reveal different clues. Even a path that gets stuck can show what your reasoning missed. Ideas and equations emerge through these attempts, questions, and deductions. Let us discover how they connect.
-
The model is trained. Now the user wants a dog in a red hat.
Predict, plot, and interpret evidence within each complete method route, then derive the details. Review the table last and reason about combined methods and new requirements.
Start with the situation → Skill manual: The Model Is Trained. Can We Still Change What We Want? -
Before the image is finished, how do we remind it to draw a corgi?
Run classifier guidance and CFG on the same mixture. Read both complete routes, then compare prerequisites and interpret new scenarios.
Start with the situation → Skill manual: Where Can a Fixed Model Get a Guidance Direction? -
Both photos look plausible. Which should we draw more often?
Move from preferred-photo proportions to continuous sampling. Plot posteriors, mean estimates, and h_t to inspect uncertainty lost by plug-in.
Start with the situation → Skill manual: High Reward and Plausible Data: Which Distribution Should We Sample? -
The evaluator only gives scores. How should these drafts continue?
Explore completed-output selection, intermediate SMC, and source MCMC/SPT in your chosen order. Sample, trace ancestry and chains, then build the final comparison.
Start with the situation → Skill manual: What If the Reward Scores Samples but Gives No Direction? (Chinese)