```xml <result> <column_name>tokens_in</column_name> <family>additive</family> <subtype>ordinal_additive</subtype> <confidence>0.85</confidence> <requires_user_confirmation>false</requires_user_confirmation> <sum_meaningful>true</sum_meaningful> <signed>false</signed> <reducers> <primary_reducer>sum</primary_reducer> <valid_reducers>sum, mean, min, max, median, count, distribution</valid_reducers> <excluded_reducers>product (not a ratio), geomean (not a ratio)</excluded_reducers> <average_semantics>arithmetic_mean</average_semantics> </reducers> <inner_outer_measure> <inner_measure_reducer>sum</inner_measure_reducer> <outer_measure_reducer>sum</outer_measure_reducer> </inner_outer_measure> <multiplicative_partner_derivable>true</multiplicative_partner_derivable> <reasoning> Gate prior: PATH table with sequence axis allows both multiplicative and ordinal_additive. Column name "tokens_in" indicates a count/quantity (additive signal). Statistics show large unbounded magnitude (max 380164, mean 5687), strictly non-negative with zeros present, and center far from 1.0 - all strongly additive. The ordinal_additive subtype is chosen over real_additive because this is a PATH table measuring an instantaneous rate/flow at each sequence step (tokens consumed per step), not a cumulative total. Summing across entities at a fixed sequence step is meaningful; along the sequence axis, normalized reducers would apply. </reasoning> <table_type_conflict>false</table_type_conflict> </result> ```