HyLaT proposes a hybrid latent-text communication protocol with two-stage training that reduces overhead while maintaining performance in multi-agent LLM systems.
Method In-Domain Out-of-Domain InferenceEfficiencyCommonsense StrategyQA SocialIQA WorldTree PubMedQAMedQA ARC-E ARC-CAvg
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
HyLaT: Efficient Multi-Agent Communication via Hybrid Latent-Text Protocol
HyLaT proposes a hybrid latent-text communication protocol with two-stage training that reduces overhead while maintaining performance in multi-agent LLM systems.