BEHAVIOR LABGame classification review
Browse mode

Which games bring out the behavior?

Three annotators review the same sample independently, in different orders. Do not discuss answers until all three finish.

TARGET BEHAVIOR

01 / CLASSIFIER INPUT

Rules & reference

Full original catalog input & implementation metadata
02 / CLASSIFIER OUTPUT

Cited evidence fields
Exact classifier output
Classifier instructions

03 / HUMAN ASSESSMENT

Your review

One required question: does the selection verdict make sense for this behavior? The rest is optional. Judge this exact game variant against the target definition. This is unblinded expert review, not observed gameplay.

What counts as a fitting verdict?

Answer Yes if the classifier’s category is appropriate: direct relevance needs no change; parameter variation uses existing settings; rule/engine mutation changes mechanics. For “not relevant” or “insufficient evidence”, Yes means you agree with that negative or uncertain classification—not that the game elicits the behavior.

Judge proposed changes separately: are they concrete, feasible, and likely to create an opportunity for the specified behavior while preserving a recognizable version of the game? New rule explanations below the original input are reviewer context, not extra information given to the classifier.

If you cannot judge, save a draft or skip with a reason rather than guessing Yes or No. Skips remain missing data, not negative labels. Work independently and retain your private resume link.

Optional explanation why
Start a reviewer session above to save annotations.