Scorio evaluates repeated outcomes, ranks models, and selects sampled answers.
This entry point exposes the evaluation metrics under the eval namespace, mirroring import scorio; scorio.eval.bayes(...) in Python:
eval
import scorio; scorio.eval.bayes(...)
import { eval as scorioEval } from "scorio";const [mu, sigma] = scorioEval.bayes([0, 1, 1]); Copy
import { eval as scorioEval } from "scorio";const [mu, sigma] = scorioEval.bayes([0, 1, 1]);
The metrics are also importable directly from scorio/eval.
scorio/eval
Ranking estimators live under the rank namespace (scorio.rank.borda(...)) and are also importable directly from scorio/rank.
rank
scorio.rank.borda(...)
scorio/rank
Sequential-inference helpers live under the sinf namespace (scorio.sinf.shouldStop(...)) and are also importable from scorio/sinf.
sinf
scorio.sinf.shouldStop(...)
scorio/sinf
Test-time-scaling aggregation lives under aggregate (with short alias agg) and is also importable directly from scorio/aggregate.
aggregate
agg
scorio/aggregate
Ranking comparison and collision-free ranking hashes live under utils and are also importable directly from scorio/utils.
utils
scorio/utils
Scorio evaluates repeated outcomes, ranks models, and selects sampled answers.
This entry point exposes the evaluation metrics under the
evalnamespace, mirroringimport scorio; scorio.eval.bayes(...)in Python:The metrics are also importable directly from
scorio/eval.Ranking estimators live under the
ranknamespace (scorio.rank.borda(...)) and are also importable directly fromscorio/rank.Sequential-inference helpers live under the
sinfnamespace (scorio.sinf.shouldStop(...)) and are also importable fromscorio/sinf.Test-time-scaling aggregation lives under
aggregate(with short aliasagg) and is also importable directly fromscorio/aggregate.Ranking comparison and collision-free ranking hashes live under
utilsand are also importable directly fromscorio/utils.