Kai: the invariance suite asks both noul swaps, and cardinality prose cites no label space the Rust port cannot reproduce.

Co-authored-by: Hanzo Dev <[email protected]>

kai/sections/08-evaluation.tex

@@ -80,11 +80,9 @@ kept runs are at \texttt{4e57fb2}.
     its concept among co-hyponyms and filler). An unanswered question counts
     as wrong. Laya rejects a question whose option markers fall past its
     context; Jev refuses more than 255 options. Laya's authors' recipe for many
-    options is a top-20 shortlist; with a general-purpose bi-encoder as the
-    shortlister, the shortlist holds the gold 66\%, 40\% and 24\% of the time at
-    1{,}000, 10{,}000 and 100{,}000 options, which bounds any model answering
-    from it, and Laya reaches 0.43, 0.20 and 0.06 and Jev 0.65, 0.40 and 0.24
-    \meas.
+    options is a top-20 shortlist from an encoder, which caps accuracy at the
+    shortlist's recall; the table reports every model on the full option set,
+    where Kai's retrieval keeps the mass it cuts (Section~\ref{sec:factor}).
   \item[Joint coherence.] Two families of synthetic programs, 200 states each,
     ask five dependent questions of one state (a vehicle's fault, severity,
     readiness, action and review; a source selection's option, evidence, risk,
@@ -97,8 +95,10 @@ kept runs are at \texttt{4e57fb2}.
   \item[Invariance.] 200 Banking77 questions of 16 options asked in five
     orders: the share whose top answer changes and the largest probability
     change; 400 \op{noul} questions asked with the two sides' descriptions
-    exchanged: the share whose answer mirrors. Proposition~\ref{prop:symmetry}
-    predicts no flips and exact mirrors for Kai, up to rounding.
+    exchanged, and again with the statement negated: the share whose answer
+    mirrors. Proposition~\ref{prop:symmetry} predicts no flips and exact
+    mirrors under the exchange for Kai, up to rounding; under negation a
+    mirror is learned, not guaranteed, for every backend.
   \item[Media.] Human activity from smartphone inertial signals, a turbofan's
     remaining life from C-MAPSS telemetry, and a valve's condition from its
     sound (MIMII): the baselines read text renderings of the signals, Kai the