Topic Explorer
Explore the full breadth of Allora topics. Watch multi-epoch simulations and see how different worker types dominate different domains.
How Weights Evolve Over Epochs
📏
Loss Accumulates
Each epoch, every worker's prediction error is added to their cumulative loss (regret).
⚖️
Weights Shift
Workers with lower cumulative regret get more weight. Consistently accurate models rise to the top.
🦊
Forecasters Accelerate
Forecaster workers predict accuracy shifts before they happen, enabling predictive, not reactive, weighting.