# GateRQ cuts conditional-memory table use in two LLMs

> A new study proposes choosing how many bits to read for each memory lookup, reporting lower GPU table memory and higher vLLM throughput.

Oossa · 2026-10-28 · https://oossa.com/en/gaterq-cuts-conditional-memory-table-use-in-two-llms

Researchers introduce GateRQ, a method for compressing conditional-memory tables—large stores of text patterns attached to a language model. Instead of assigning extra precision to frequently accessed rows, GateRQ uses a lookup’s gate value to decide whether to read additional 1-bit codes; the paper says the gate can be computed from a 1-bit code.

On DeepSeek-V4.1-Flash, the researchers report 7.8 times less GPU table memory and 11%–31% higher vLLM throughput than with FP8 tables. They report losing about 1 PopQA point at an average 1.50 bits read per lookup, or 0.2 points at 2.10 bits, across two models; the method does not require retraining.

## The facts

- The paper is arXiv:2610.35985, posted on October 28, 2026.
- The researchers report 7.8× lower GPU table memory on DeepSeek-V4.1-Flash.

## Why it matters

For teams serving models with conditional-memory tables, GateRQ could reduce GPU memory use while increasing throughput, though the reported results do not establish how it performs outside the tested models.

## Sources & references

1. [Let the Gate Decide: Per-Lookup Precision for Quantized Conditional Memory in LLMs](https://arxiv.org/abs/2610.35985) – arXiv cs.LG, 2026-10-28

Last updated: 2026-10-28