external benchmark
MysteryMechanism leaderboard
Vals MysteryMechanism. Every model the catalog carries a published MysteryMechanism value for, ranked by that value.
Categoryexternal
MeasureAccuracy score
TasksSealed mathematical mechanisms probed under an experiment budget
DifficultyAgentic scientific discovery
Published byVals MysteryMechanism
Can agents rediscover sealed mathematical mechanisms through bounded experiments?
No model in the catalog has a published MysteryMechanism score.
The benchmark is defined by Vals MysteryMechanism, but the catalog carries no value for it yet. An absent value is shown as absent here rather than as a zero.