Every hidden oracle ships a seeded validation set (one or more independently written correct solutions whose replication script is re-run, and two to six wrong solutions). The table lists the outcome of \texttt{evaluators/run\_validations.py} executed before the freeze: PASS means every seeded case received the expected verdict (no false acceptance and no false rejection on the seeded set). The seeded sets bound gross evaluator errors; they do not cover the space of valid solutions. \begin{longtable}{p{1.0in}p{0.6in}p{4.6in}} \toprule Task & Result & Last line of the validation output \\ \midrule \endhead F1-conf-01 & PASS & ALL PASS \\ F1-conf-02 & PASS & ALL PASS \\ F1-conf-03 & PASS & ALL PASS \\ F1-conf-04 & PASS & ALL PASS \\ F1-dev-01 & PASS & ALL PASS \\ F2-conf-01 & PASS & ALL OK \\ F2-conf-02 & PASS & ALL OK \\ F2-conf-03 & PASS & ALL OK \\ F2-conf-04 & PASS & ALL OK \\ F2-dev-01 & PASS & ALL OK \\ F2-smoke-01 & no validate.py & \\ F3-conf-01 & PASS & ALL PASS \\ F3-conf-02 & PASS & ALL PASS \\ F3-conf-03 & PASS & ALL PASS \\ F3-conf-04 & PASS & ALL PASS \\ F3-dev-01 & PASS & ALL PASS \\ F4-conf-01 & PASS & VALIDATION PASSED \\ F4-conf-02 & PASS & VALIDATION PASSED \\ F4-conf-03 & PASS & VALIDATION PASSED \\ F4-conf-04 & PASS & VALIDATION PASSED \\ F4-dev-01 & PASS & VALIDATION PASSED \\ F4-sources & no validate.py & \\ F5-conf-01 & PASS & ALL PASS \\ F5-conf-02 & PASS & ALL PASS \\ F5-conf-03 & PASS & ALL PASS \\ F5-conf-04 & PASS & ALL PASS \\ F5-dev-01 & PASS & ALL PASS \\ F6-conf-01 & PASS & VALIDATION PASSED \\ F6-conf-02 & PASS & VALIDATION PASSED \\ F6-conf-03 & PASS & VALIDATION PASSED \\ F6-conf-04 & PASS & VALIDATION PASSED \\ F6-dev-01 & PASS & VALIDATION PASSED \\ \bottomrule \end{longtable}