diff --git a/main.tex b/main.tex
index 9ff0d82..01e487f 100644
--- a/main.tex
+++ b/main.tex
@@ -154,8 +154,14 @@ pipeline proceeds as follows:
     previously kept sample $\mathbf{s}'$.
 \end{enumerate}
 
-An optional LLM-as-Judge scoring module provides multi-dimensional
-quality scores that can be used to filter low-quality samples.
+In addition to deduplication, instruction--response pairs whose
+response contains fewer than 15 tokens are discarded.
+Short replies exhibit high per-token variance in both
+conditional and unconditional loss (Appendix~\ref{sec:ifd_bias},
+Figure~\ref{fig:length_bias}), which would otherwise distort
+the IFD-based difficulty estimates used for downstream sample
+selection.  This length floor is applied per-field, analogous to
+the pretraining filter described above.
 
 \subsection{DPO Data Generation}