kvleverage

null resultshipPython · PyTorchModel behavior & reasoningJuly 12, 2026

the question

Does the attention-sink token span the key basis of the KV cache, justifying anchor-token compression?

what came out

No. Leverage concentration of the real cache is at or below a matched-rank random cache (max/mean 1.12 to 1.46), and the sink is essentially never a top-5 leverage position except for 0.5B keys. KV rank is delocalized; the pre-registered spanning-anchor hypothesis is falsified. Attention mass and basis leverage are different things.

method & receipts

Result: null result — an honest negative, reported as it came out.
Pre-registered — the prediction was written down and committed before the run.
Reproducible — one script re-runs the whole thing from scratch.
Tested — a correctness/benchmark suite ships alongside the code.

→ read the code and re-run it

github.com/v-code01/kvleverage