Fine-Tuning a KV Cache Concatenation-Aware Model or Recomputing KV Caches? Why Not Both?

Experiments on the RULER benchmark show that, for a 124k-token input, our method improves the RULER score by 9.7 point over the baseline that recomputes KV caches only. It comes from a paper posted to arXiv on 10 September 2026. In Retrieval-Augmented Generation (RAG) systems, a large number of retrieved chunks are concatenated to form the input context so that users can receive high-quality responses based on external knowledge. As a result, the input context length increases substantially, leading to a larger prefill workload and, in turn, a longer time to first token (TTFT). While previous works that reuse precomputed key-value (KV) caches effectively reduce TTFT for long-context inputs, it remains unclear whether response quality is preserved when the input context becomes very long. In this paper, we propose a combined approach that (i) fine-tunes the model while taking KV cache concatenation into account and (ii) selectively recomputes a subset of the KV caches. By applying both techniques, we demonstrate improved accuracy for long-context inputs. Moreover, TTFT is reduced by 80% compared with full attention.

Why it counts

Experiments on the RULER benchmark show that, for a 124k-token input, our method improves the RULER score by 9.7 point over the baseline that recomputes KV caches only. By applying both techniques, we demonstrate improved accuracy for long-context inputs. Moreover, TTFT is reduced by 80% compared with full attention.

Sources

Primary source: primary source

What is not known

This brief does not claim independent replication. Claims that appear only on X and not in the primary source stay unknown.