The Sequential Edge: Inverse-Entropy Voting Beats Parallel Self-Consistency at Matched Compute
This work challenges the dominant parallel reasoning paradigm in LLM inference, potentially impacting researchers and practitioners in AI and NLP by offering a more effective test-time scaling strategy.
The paper tackles the problem of test-time scaling for language model reasoning by comparing parallel self-consistency with sequential refinement, finding that sequential scaling outperforms parallel in 95.6% of configurations with accuracy gains up to 46.7% and introduces inverse-entropy weighted voting to further boost accuracy.
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.