---
title: "Verifier-Guided Self-Consistency: Test-Time Reasoning with Imperfect Generators and Verifiers"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2611.06606v1"
permalink: "https://feed.astralor.com/items/itm_99162602aa4e4e583fb9845b"
published: "2026-11-04T14:57:04.000Z"
score: 59
channel: "tech"
tags: ["LLM","RAG","Verifier-Guided Self-Consistency","Test-Time Reasoning","Imperfect Generators","Verifiers Large","Improving","Large"]
---

# Verifier-Guided Self-Consistency: Test-Time Reasoning with Imperfect Generators and Verifiers

## At a glance

Large language models can solve many multi-step reasoning problems, but their reasoning remains brittle: plausible intermediate steps often lead to incorrect answers. Improving reasoning at test time requires balancing …

## Key points

- Large language models can solve many multi-step reasoning problems, but their reasoning remains brittle: plausible intermediate steps often lead to incorrect answers.
- Improving reasoning at test time requires balancing two imperfect signals: the generator, which provides diversity, and the verifier, which guides correctness.
- Existing approaches largely favor one side of this tradeoff.

## Full text

Large language models can solve many multi-step reasoning problems, but their reasoning remains brittle: plausible intermediate steps often lead to incorrect answers. Improving reasoning at test time requires balancing two imperfect signals: the generator, which provides diversity, and the verifier, which guides correctness. Existing approaches largely favor one side of this tradeoff. Verifier-weighted voting relies on trajectories sampled from the generator and cannot steer generation toward correct ones. By contrast, beam-style verifier-guided search can steer generation, but may over-concentrate on narrow trajectory sets and amplify verifier errors. We propose Verifier-Guided Self-Consistency (VGSC), a stochastic decoding framework that combines step-level correctness and future-outcome estimates to guide reasoning during generation. VGSC repeatedly proposes candidate next steps from the base language model and uses both signals in an acceptance/resampling procedure, while maintaining multiple trajectories for final answer-level self-consistency. We provide a probabilistic interpretation of VGSC as approximate sampling from a verifier-tilted trajectory distribution. Across multiple reasoning benchmarks, generator models, and verifier families, VGSC achieves higher average accuracy than representative baselines.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_99162602aa4e4e583fb9845b