Saved in:
Bibliographic Details
Main Authors: Bao, Wenrui, Wang, Huan, Wang, Jian, Wang, Zhangyang, Wang, Kai, Shang, Yuzhang
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.13839
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916010346414080
author Bao, Wenrui
Wang, Huan
Wang, Jian
Wang, Zhangyang
Wang, Kai
Shang, Yuzhang
author_facet Bao, Wenrui
Wang, Huan
Wang, Jian
Wang, Zhangyang
Wang, Kai
Shang, Yuzhang
contents Multi-agent LLM systems usually collaborate by exchanging natural-language messages. This interface is simple and interpretable, but it forces each sender's intermediate computation to be serialized into tokens and then reprocessed by the receiver, thereby increasing the generated-token cost, prefill overhead, and KV-cache memory. We study an alternative communication interface: instead of appending a sender's message to the receiver's context, compile the sender's hidden states into a transient, receiver-specific weight perturbation. We introduce TFlow (Thought Flow), a weight-space communication framework for a known and fixed receiver architecture. For each query, frozen role-prompted sender agents process the input, and a learned parameter generator maps their internal activations into low-rank LoRA perturbations targeting the receiver's modules. These perturbations are fused and applied only during the receiver's generation, enabling instance-level adaptation without permanently changing the model or enlarging the receiver's text context. With three Qwen3-4B agents, TFlow improves over a standalone receiver by up to 8.5 accuracy points across five benchmarks while reducing processed tokens by up to 32.69%. Compared with a text-based three-agent baseline, it reduces total processed tokens by up to 83.27% and the wall-clock inference time by up to 4.6$\times$, while maintaining competitive accuracy on four of five benchmarks. These results suggest that transient low-rank weight perturbations can serve as an executable communication medium for efficient multi-agent LLM collaboration.
format Preprint
id arxiv_https___arxiv_org_abs_2605_13839
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Good Agentic Friends Do Not Just Give Verbal Advice: They Can Update Your Weights
Bao, Wenrui
Wang, Huan
Wang, Jian
Wang, Zhangyang
Wang, Kai
Shang, Yuzhang
Computation and Language
Multi-agent LLM systems usually collaborate by exchanging natural-language messages. This interface is simple and interpretable, but it forces each sender's intermediate computation to be serialized into tokens and then reprocessed by the receiver, thereby increasing the generated-token cost, prefill overhead, and KV-cache memory. We study an alternative communication interface: instead of appending a sender's message to the receiver's context, compile the sender's hidden states into a transient, receiver-specific weight perturbation. We introduce TFlow (Thought Flow), a weight-space communication framework for a known and fixed receiver architecture. For each query, frozen role-prompted sender agents process the input, and a learned parameter generator maps their internal activations into low-rank LoRA perturbations targeting the receiver's modules. These perturbations are fused and applied only during the receiver's generation, enabling instance-level adaptation without permanently changing the model or enlarging the receiver's text context. With three Qwen3-4B agents, TFlow improves over a standalone receiver by up to 8.5 accuracy points across five benchmarks while reducing processed tokens by up to 32.69%. Compared with a text-based three-agent baseline, it reduces total processed tokens by up to 83.27% and the wall-clock inference time by up to 4.6$\times$, while maintaining competitive accuracy on four of five benchmarks. These results suggest that transient low-rank weight perturbations can serve as an executable communication medium for efficient multi-agent LLM collaboration.
title Good Agentic Friends Do Not Just Give Verbal Advice: They Can Update Your Weights
topic Computation and Language
url https://arxiv.org/abs/2605.13839