search

synapse-alpha / mirror-neuron

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

Pearson correlation scores versus rewards for each time step #55

Closed steffencruz closed 1 year ago

steffencruz commented 1 year ago

Also use spearman coeff

joeylegere commented 1 year ago

https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.pearsonr.html

joeylegere commented 1 year ago

maybe try: https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.kendalltau.html