# Latent Communication

4 related articles

Subscribe to updates

arXiv Nov 4, 2026
Tool outputs no longer have to be text: Coca-Cola compresses search results into latent tokens, cutting a web-search agent's token use by more than 70%
Each time a web-search agent calls a tool, the long returned text is appended verbatim to its context, so after several steps the context balloons and dilutes attention. Coca-Cola instead compresses those observations into a few continuous latent tokens and embeds them directly in the agent's input embedding space, with action-preserving training aligning the next-action distributions under full and compressed observations. Across four QA benchmarks and two model families, total token consumption falls by 70.01%–89.24% and end-to-end runtime by 24.45%–55.97%, while EM and F1 exceed the uncompressed agent in every setting.
Each time a web-search agent calls a tool, the long returned text is appended verbatim to its context, so after several steps the context balloons and dilutes attention. Coca-Cola instead compresses those observations into a few continuous latent tokens and embeds them directly in the agent's input embedding space, with action-preserving training aligning the next-action distributions under full and compressed observations. Across four QA benchmarks and two model families, total token consumption falls by 70.01%–89.24% and end-to-end runtime by 24.45%–55.97%, while EM and F1 exceed the uncompressed agent in every setting.
Tsinghua University
Yuetong Liu（First author）
Coca-Cola Context compression LLM Agents Latent Communication
arXiv Oct 21, 2026
Diffusion-LM agents skip the text step: one hidden message beats four exchanges by 7.0 points on QASC
Diffusion language model agents can pass compressed denoising states to each other without decoding text first. A sender packs its hidden state into four signed dense vectors, and the receiver keeps its own context and trajectory while denoising. On QASC, matched messages raise accuracy from 72.6% to 80.8%, a message from another question drops it to 46.8%, and one late transfer reaches 87.8%, 7.0 points above four recurrent exchanges.
Diffusion language model agents can pass compressed denoising states to each other without decoding text first. A sender packs its hidden state into four signed dense vectors, and the receiver keeps its own context and trajectory while denoising. On QASC, matched messages raise accuracy from 72.6% to 80.8%, a message from another question drops it to 46.8%, and one late transfer reaches 87.8%, 7.0 points above four recurrent exchanges.
ITMO University National Research University Higher School of Economics
Alan-Barsag Gazzaev（First author）
Latent Communication Diffusion Language Model multi-agent systems
arXiv Oct 1, 2026
Freezing agents and training only latent links: benign link training alone raises harmful compliance above text communication, and an RL attack lifts the mean from 27.9 to 76.9
This work studies the safety of latent communication in multi-agent systems, where lightweight trainable links replace text messages, and finds that with all agent parameters frozen, benign link training alone increases harmful compliance relative to text communication, that an attacker can amplify this via supervised optimization on harmful query-response pairs, poisoning benign training data, or a reward-guided reinforcement-learning attack that needs no harmful target responses, raising the mean harmful-compliance score from 27.9 with benignly trained links to 76.9 across three topologies and four safety benchmarks, and that adapting the rewards toward safer behavior repairs compromised links by updating only the links.
This work studies the safety of latent communication in multi-agent systems, where lightweight trainable links replace text messages, and finds that with all agent parameters frozen, benign link training alone increases harmful compliance relative to text communication, that an attacker can amplify this via supervised optimization on harmful query-response pairs, poisoning benign training data, or a reward-guided reinforcement-learning attack that needs no harmful target responses, raising the mean harmful-compliance score from 27.9 with benignly trained links to 76.9 across three topologies and four safety benchmarks, and that adapting the rewards toward safer behavior repairs compromised links by updating only the links.
Muhammad Huzaifa（First author）
Latent Communication Safety Alignment GRPO Reward-guided link repair LLM-Based Multi-Agent Systems
arXiv Sep 29, 2026
Draft-KV makes a receiver actually use the sharer's draft KV states: a 0.5B receiver rises from 37.45% to 78.04% on MMLU-Redux, and falls back to 36.40% when the message comes from an unrelated question
The work first audits existing latent communication interfaces: across five method-dataset pairs, replacing each message with one from an unrelated question changes accuracy by at most 0.60 points even when communication adds 15.44 points over the receiver alone, indicating the gain can come from interface adaptation rather than message content; it then introduces Draft-KV, which sends the key-value states formed while the sharer drafts an answer through linear projections into a side memory read by a gated attention branch, trained progressively from message reconstruction to answer supervision under a one-sided guard on harm from mismatched messages, with both models frozen and only 1.05M interface parameters trained (348x fewer than C2C); with a Qwen3-8B sharer and a frozen Qwen2.5-0.
The work first audits existing latent communication interfaces: across five method-dataset pairs, replacing each message with one from an unrelated question changes accuracy by at most 0.60 points even when communication adds 15.44 points over the receiver alone, indicating the gain can come from interface adaptation rather than message content; it then introduces Draft-KV, which sends the key-value states formed while the sharer drafts an answer through linear projections into a side memory read by a gated attention branch, trained progressively from message reconstruction to answer supervision under a one-sided guard on harm from mismatched messages, with both models frozen and only 1.05M interface parameters trained (348x fewer than C2C); with a Qwen3-8B sharer and a frozen Qwen2.5-0.
Huawei Technologies (China)Tencent (China)Hong Kong Metropolitan University City University of Hong Kong University of Electronic Science and Technology of China
Linquan Wu（First author）
Draft-KV Latent Communication KV Cache Model Collaboration

You’re all caught up

Tool outputs no longer have to be text: Coca-Cola compresses search results into latent tokens, cutting a web-search agent's token use by more than 70%

Diffusion-LM agents skip the text step: one hidden message beats four exchanges by 7.0 points on QASC

Freezing agents and training only latent links: benign link training alone raises harmful compliance above text communication, and an RL attack lifts the mean from 27.9 to 76.9

Draft-KV makes a receiver actually use the sharer's draft KV states: a 0.5B receiver rises from 37.45% to 78.04% on MMLU-Redux, and falls back to 36.40% when the message comes from an unrelated question