pyramids/simple-v0: re-record with the action-label gap fix (2.98% of rows held a placeholder 0); band U(4.8,7.1) unchanged

pyramids/medium-v0: re-record with the action-label gap fix (4.46% of rows held a placeholder 0); band U(2.7,4.9) unchanged

pyramids/expert-v0: fix action labels — the collector recorded a fabricated 0 for every step an agent spent inside ML-Agents' action-repeat gap (~6% of rows); it now carries the last real decision forward, so the action column matches the policy the trajectories were actually driven by

pyramids/simple-v0: fixed-T -> per-episode softmax-temperature-range tier (T~U(4.8,7.1) per episode, step-reward norm 0.6011)

pyramids/medium-v0: fixed-T -> per-episode softmax-temperature-range tier (T~U(2.7,4.9) per episode, step-reward norm 0.8076)

pyramids/simple-v0: replace epsilon tier with softmax-temperature tier (T=6.0, step-reward calibrated)

pyramids/medium-v0: replace epsilon tier with softmax-temperature tier (T=3.7, step-reward calibrated)

feat(pyramids): re-collect simple with per-episode epsilon pool (trajectory diversity)

feat(pyramids): re-collect medium with per-episode epsilon pool (trajectory diversity)

pyramids/simple re-collect to normalized ~0.60 (eps 0.67, return 0.6670)

add pyramids/medium-v0 (Minari-aligned ~0.79, epsilon=0.59)

remove pyramids/medium-v0 (relabeled to simple; new 0.75 medium to be re-collected)

relabel pyramids medium->simple (Minari-aligned tiers): copy hdf5, rewrite metadata/sidecars

Add unity/pyramids/simple-v0 (stock Pyramids.onnx + epsilon=0.77 action noise; normalized return ~0.27)

Add unity/pyramids/medium-v0 (stock Pyramids.onnx + epsilon=0.70 action noise; normalized return ~0.51)

Add Pyramids release-23 expert trajectories