← 研踪 PaperTrailarXiv:2605.06219 · v2
Joint Consistency: A Unified Test-Time Aggregation Framework via Energy Minimization
首次提交 2026-05-07 21:17
版本更新 2026-11-15 12:47
雷达观测 2026-11-17 19:15
摘要
This paper studies test-time aggregation, an approach that generates multiple reasoning traces and aggregates them into a final answer. Most existing methods rely on evaluation signals collected from candidate traces in isolation or answer frequencies, while ignoring comparative interactions among candidates. We propose Joint Consistency (JC), formulated as a constrained Ising-type energy minimization problem, which employs both independent evaluation signals and pairwise comparisons using LLM-as-a-judge. JC provides a unified framework for test-time aggregation that subsumes typical voting and weighted aggregation methods as special cases. Moreover, we develop an efficient approximation strategy that makes interaction modeling practical for large-scale test-time aggregation. Experiments on math and code reasoning benchmarks show that JC outperforms existing baselines across tasks, judge models, trace budgets, and trace-generation settings, with particularly pronounced gains on challenging problems.
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v2 · 2026-11-15 12:47 · 最新观测版本
Joint Consistency: A Unified Test-Time Aggregation Framework via Energy Minimization
This paper studies test-time aggregation, an approach that generates multiple reasoning traces and aggregates them into a final answer. Most existing methods rely on evaluation signals collected from candidate traces in isolation or answer frequencies, while ignoring comparative interactions among candidates. We propose Joint Consistency (JC), formulated as a constrained Ising-type energy minimization problem, which employs both independent evaluation signals and pairwise comparisons using LLM-as-a-judge. JC provides a unified framework for test-time aggregation that subsumes typical voting and weighted aggregation methods as special cases. Moreover, we develop an efficient approximation strategy that makes interaction modeling practical for large-scale test-time aggregation. Experiments on math and code reasoning benchmarks show that JC outperforms existing baselines across tasks, judge models, trace budgets, and trace-generation settings, with particularly pronounced gains on challenging problems.
查看此版本解读 →查看此版本原文 ↗