TwinKV repairs KV cache eviction using key redundancy, not attention

A controlled leave-one-out probe finds that attention magnitude is unrelated to a token's causal contribution to the answer (Spearman rho = -0.004), undermining the premise behind dominant attention-based KV eviction. TwinKV is a training-free repair pass that composes on top of any eviction policy: it detects evicted tokens whose keys have no surviving near-duplicate (orphans) and retained tokens whose information is duplicated elsewhere (redundant donors), then swaps them while preserving the original budget and scoring rule. Composed with four recent policies on Qwen3-4B across LongBench, LooGLE, RULER and an MMLU-Pro no-harm control at compression ratios 0.3/0.5/0.7, it improves a majority of configurations for two policies, is near-even for a third, and helps only a minority for an adaptive baseline already near its performance ceiling.

The composable post-pass design means serving teams can trial TwinKV on top of existing eviction configs without retraining or replacing their policy — rare for KV-cache research. The null result on attention-as-importance is a separate, useful caution for anyone tuning eviction thresholds by hand.