VeriLLM: A Lightweight Framework for Publicly Verifiable Decentralized Inference

2025-09-29

AI Summary

In decentralized LLM inference, permissionless nodes and absence of prior trust render outputs unverifiable. Method: We propose the first lightweight decentralized inference verification framework with game-theoretic security guarantees. Our approach features a homogeneous inference-verification network architecture enabling GPU resource reuse; introduces an LLM-optimized lightweight verification algorithm and a public verification protocol under the one-honest-verifier assumption; and designs a peer-prediction-based incentive mechanism ensuring task indistinguishability and Nash equilibrium for honest behavior. Contributions/Results: Experiments show verification overhead reduced to ~1% of inference cost, significant end-to-end throughput improvement, and effective expansion of the verifiable node pool. We formally prove system security and stability against rational adversaries.

Technology Category

• Multiagent Systems: Mechanism Design
• Machine Learning: Large Multimodal Models (LMMs)
• Reasoning under Uncertainty: Sequential Decision Making

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Decentralized Web and Fediverse systems

Abstract

Decentralized inference is an appealing paradigm for serving large language models (LLMs), offering strong security, high efficiency, and lower operating costs. Yet the permissionless setting admits no a priori trust in participating nodes, making output verifiability a prerequisite for secure deployment. We present VeriLLM, a publicly verifiable protocol for decentralized LLM inference that (i) achieves security under a one-honest-verifier assumption, (ii) attains near-negligible verification cost (about 1% of the underlying inference) via a lightweight verification algorithm designed explicitly for LLMs, and (iii) enforces honest checking through a peer-prediction mechanism that mitigates lazy verification in naive voting. We further introduce an isomorphic inference-verification network that multiplexes both roles on the same set of GPU workers. This architecture (i) increases GPU utilization and thereby improves end-to-end throughput for both inference and verification, (ii) expands the effective pool of available validators, strengthening robustness and security, and (iii) enforces task indistinguishability at the worker boundary to prevent job-type-conditioned behavior. Finally, we provide a formal game-theoretic analysis and prove that, under our incentives, honest inference and verification constitute a Nash equilibrium, ensuring incentive compatibility against rational adversaries. To our knowledge, this is the first decentralized inference verification protocol with an end-to-end game-theoretic security proof.

Problem

• Ensuring verifiable outputs in permissionless decentralized LLM inference
• Reducing verification costs through lightweight algorithms for LLMs
• Establishing game-theoretic security against rational adversaries in decentralized systems

Innovation

• Lightweight verification algorithm for LLMs
• Isomorphic inference-verification network architecture
• Game-theoretic Nash equilibrium for security