Hermes SkillEval Dashboard

Data provenance

Data source
docs/demo/phase7b-cross-encoder-calibration
Generated from
4 test-split runs
Tasks
30 held-out test records
Generated by
skilleval dashboard
Generated at
2026-05-25T09:03:13+00:00
Input runs
cross-encoder-calibrated-balanced-test, cross-encoder-calibrated-strict-test, cross-encoder-rank-only-test, gated-minilm-contrastive-test
Git commit
689a97d
Run Task Split Category Difficulty Recall@5 Neg Hit Abstain Selected Tags