Evaluation Run 20261015T005906

Rubric 2026-09-23b · typesafe/jev-1.13 · 2 of 2 answered · 0.2s · $0.0002

No entry file was written. Results live in this run's ledger until Phase 4 adds the write path, after calibration sets the thresholds.

Per-dimension distribution

DimensionTypeDistributionMean confidenceAgrees
answer_key_correctnoulmean 0.98 · level 1 holds 2/2 FLAT——
difficultyscoremean 0.45 · level 0 holds 2/2 FLAT0.562/2
distractor_pullscoremean 1.10 · level 1 holds 2/2 FLAT0.48—
engagementscoremean 0.98 · level 1 holds 2/2 FLAT0.70—
one_defensible_answernoulmean 0.90 · level 1 holds 2/2 FLAT——
options_scan_cleanscoremean 1.75 · level 2 holds 2/2 FLAT0.61—
passage_realismscoremean 1.17 · level 1 holds 2/2 FLAT0.49—
reads_in_one_passscoremean 1.92 · level 2 holds 2/2 FLAT0.88—
sat_choice_conventionsscoremean 1.98 · level 2 holds 2/2 FLAT0.98—
sat_stem_conventionsscoremean 1.81 · level 2 holds 2/2 FLAT0.70—
sectionchoicewriting 2/21.002/2
single_movenoulmean 0.94 · level 1 holds 2/2 FLAT——
skill_codechoicestandard-english 2/20.922/2
stem_self_containednoulmean 0.96 · level 1 holds 2/2 FLAT——
survives_speednoulmean 0.84 · level 1 holds 2/2 FLAT——
tested_construct_is_satnoulmean 0.97 · level 1 holds 2/2 FLAT——

A dimension marked FLAT holds more than 80% of the run on one level. That is either a clean bank or a dimension reporting a constant, and only a negative control tells them apart.

Run

Served by{'typesafe/jev-1.13-20260917 via TypeSafe': 2}
Input tokens4,592
Failed0
Preflight notes0
Skipped entries0