Receiver-Conditioned Latent Communication gives 94% CacheBack

Hugging Face Daily Papers Papers
multi-agent kv-cache latent-communication training-free efficiency attention

Summary

The paper introduces CacheBack, a training-free receiver-conditioned latent communication method that filters and compresses an agent's KV cache based on the receiving agent's information needs, reducing transferred state by 75%, improving accuracy by 14.7 points, and cutting latency 3.2x vs text communication on FanOutQA with Qwen3.

Multi-agent systems distribute large contexts across agents that communicate to solve a task. Text messages are compact but require decoding and may omit evidence the receiving agent needs. Recent latent communication instead transfers KV caches. This avoids text generation and can improve accuracy and latency. However, a full KV cache grows linearly with both the context an individual agent processes, and the number of agents that coordinate together. This raises memory and context costs, often far exceeding available GPU resources and context window sizes. Our key observation is that agents need only send what the receiving agent requires for its local task -- which we call receiver-conditioned communication. The receiver agent passes the sender a small description of its information needs, which serves to filter and compress the sender agent's KV cache. CacheBack is a simple, robust, training-free instance of receiver conditioning based on the sender's attention weights. On FanOutQA, CacheBack with Qwen 3 removes 75% of the state the agent would otherwise receive, improving accuracy by 14.7 percentage points and reducing median task-completion latency by 3.2x relative to text communication. We show comparable improvements across model families that span dense Transformers, Mamba-attention hybrids, and sliding-window attention.
Original Article
View Cached Full Text

Cached at: 10/05/26, 04:49 PM

Paper page - Receiver-Conditioned Latent Communication gives 94% CacheBack

Source: https://huggingface.co/papers/2609.32046

Abstract

Multi-agentsystemsdistributelargecontextsacrossagentsthatcommunicatetosolveatask.Textmessagesarecompactbutrequiredecodingandmayomitevidencethereceivingagentneeds.RecentlatentcommunicationinsteadtransfersKVcaches.Thisavoidstextgenerationandcanimproveaccuracyandlatency.However,afullKVcachegrowslinearlywithboththecontextanindividualagentprocesses,andthenumberofagentsthatcoordinatetogether.Thisraisesmemoryandcontextcosts,oftenfarexceedingavailableGPUresourcesandcontextwindowsizes.Ourkeyobservationisthatagentsneedonlysendwhatthereceivingagentrequiresforitslocaltask--whichwecallreceiver-conditionedcommunication.Thereceiveragentpassesthesenderasmalldescriptionofitsinformationneeds,whichservestofilterandcompressthesenderagent’sKVcache.CacheBackisasimple,robust,training-freeinstanceofreceiverconditioningbasedonthesender’sattentionweights.OnFanOutQA,CacheBackwithQwen3removes75%ofthestatetheagentwouldotherwisereceive,improvingaccuracyby14.7percentagepointsandreducingmediantask-completionlatencyby3.2xrelativetotextcommunication.WeshowcomparableimprovementsacrossmodelfamiliesthatspandenseTransformers,Mamba-attentionhybrids,andsliding-windowattention.

View arXiv pageView PDFProject pageGitHub6Add to collection

Get this paper in your agent:

hf papers read 2609\.32046

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2609.32046 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2609.32046 in a dataset README.md to link it from this page.

Spaces citing this paper1

Collections including this paper1

Similar Articles

PorTAL (1 minute read)

TLDR AI

Introduces Latent Briefing, a method for multi-agent systems to communicate via KV cache compaction, achieving 31% fewer tokens with same accuracy and up to 20x speedup.