The Sequential Edge: Inverse-Entropy Voting Beats Parallel Self-Consistency at Matched Compute

TL;DR

This paper demonstrates that sequential reasoning with inverse-entropy voting significantly outperforms parallel self-consistency in language model inference, suggesting a paradigm shift in test-time scaling strategies.

Contribution

It introduces inverse-entropy weighted voting and establishes sequential refinement as the superior inference method over parallel self-consistency.

Findings

Sequential scaling outperforms parallel self-consistency in 95.6% of configurations.

Inverse-entropy voting boosts accuracy further over majority voting.

Sequential reasoning achieves up to 46.7% accuracy gains.

Abstract

We revisit test-time scaling for language model reasoning and ask a fundamental question: at equal token budget and compute, is it better to run multiple independent chains in parallel, or to run fewer chains that iteratively refine through sequential steps? Through comprehensive evaluation across 5 state-of-the-art open source models and 3 challenging reasoning benchmarks, we find that sequential scaling where chains explicitly build upon previous attempts consistently outperforms the dominant parallel self-consistency paradigm in 95.6% of configurations with gains in accuracy upto 46.7%. Further, we introduce inverse-entropy weighted voting, a novel training-free method to further boost the accuracy of sequential scaling. By weighing answers in proportion to the inverse entropy of their reasoning chains, we increase our success rate over parallel majority and establish it as the optimal test-time scaling strategy. Our findings fundamentally challenge the parallel reasoning orthodoxy that has dominated test-time scaling since Wang et al.'s self-consistency decoding (Wang et al., 2022), positioning sequential refinement as the robust default for modern LLM reasoning and necessitating a paradigm shift in how we approach inference-time optimization.