Skip to content

Interface: LessonScoringConfig ​

Defined in: ast/lessons-node.ts:31

Scoring and recall configuration for P.lessons.

Governs sliding-window Bayesian score calculation, partition-affinity boosting, contract-epoch decay, and top-$N$ recall filtering at session start.

Properties ​

PropertyModifierTypeDescriptionDefined in
decayOnPartitionEpochChangereadonlynumberExponential decay base $\gamma \in (0, 1]$ applied per partition contract epoch increment.ast/lessons-node.ts:37
partitionMatchBoostreadonlynumberPartition match boost factor $\beta$ applied when $L.\text{partition} \in P$.ast/lessons-node.ts:35
recallreadonlyobjectRecall filter thresholds (minScore and maxItems).ast/lessons-node.ts:39
recall.maxItemsreadonlynumber-ast/lessons-node.ts:39
recall.minScorereadonlynumber-ast/lessons-node.ts:39
slidingWindowInjectionsreadonlynumberSliding window size $K$ of deduplicated step executions used in outcome scoring.ast/lessons-node.ts:33

Released under the MIT License.