synapse-alpha / mirror-neuron

Experiments on bittensor reward models to find exploits
BSD 2-Clause "Simplified" License
1 stars 0 forks source link

Include failure rate as a configurable parameter for all models #45

Closed steffencruz closed 1 year ago

steffencruz commented 1 year ago

Reward model and gating model shouldn't be failing. We already have fail_rate in DummyDendritePool