ResearchBenchmarksLong-context Memory Evaluation on LongMemEval-S (seed 23)Follow0.05Delta ScoreRegimes0.04750.048750.050.05125Jun 8, 2026Evaluation ResultsMethodMethodLinksDelta ScoreWrite-to-Read ScoreRead-to-Write ScoreDiscordant CountConfirmation Accuracy (Base)Confirmation Accuracy (Post)Optimal Score (Base)McNemar p-value (Post-hoc)Regimespromotes (total)=1promotes (total)=12026.060.057297176720.18