InfiCoEvalChain: A Blockchain-Based Decentralized Framework for Collaborative LLM Evaluation

TL;DR

This paper introduces InfiCoEvalChain, a blockchain-based decentralized framework for LLM evaluation that enhances stability and reliability by leveraging diverse compute nodes and multi-party consensus, addressing issues of opacity and variance in traditional centralized methods.

Contribution

The paper presents a novel decentralized evaluation framework utilizing blockchain technology to improve the stability and trustworthiness of LLM benchmarking.

Findings

Decentralized evaluation reduces standard deviation from 1.67 to 0.28

Framework ensures higher statistical confidence in model rankings

Implementation is complete and will be released to the community

Abstract

The rapid advancement of large language models (LLMs) demands increasingly reliable evaluation, yet current centralized evaluation suffers from opacity, overfitting, and hardware-induced variance. Our empirical analysis reveals an alarming inconsistency in existing evaluations: the standard deviation across ten repeated runs of a single model on HumanEval (1.67) actually exceeds the performance gap among the top-10 models on the official leaderboard (0.91), rendering current rankings statistically precarious. To mitigate these instabilities, we propose a decentralized evaluation framework that enables hardware and parameter diversity through large-scale benchmarking across heterogeneous compute nodes. By leveraging the blockchain-based protocol, the framework incentivizes global contributors to act as independent validators, using a robust reward system to ensure evaluation integrity and discourage dishonest participation. This collective verification transforms evaluation from a "centralized black box" into a "decentralized endorsement" where multi-party consensus and diverse inference environments yield a more stable, representative metric. Experimental results demonstrate that the decentralized evaluation framework reduces the standard deviation across ten runs on the same model to 0.28. This significant improvement over conventional frameworks ensures higher statistical confidence in model rankings. We have completely implemented this platform and will soon release it to the community.