---
title: "共识性：无需验证器的测试时扩展新方法"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2608.09898v1"
permalink: "https://feed.astralor.com/items/itm_a874f8de9a8bec07b6a8147d"
published: "2026-08-10T17:45:44.000Z"
score: 76
channel: "ai"
tags: ["arXiv","Consilience","无验证器","测试时扩展","置信度轨迹"]
---

# 共识性：无需验证器的测试时扩展新方法

## AI 点评

> 这篇论文特别值得关注，因为它直接挑战了‘高置信度等于高正确率’的直觉，给测试时扩展提供了更精细的信号。对做推理优化或部署复杂任务智能体的团队来说，这个置信度轨迹指标非常实用，几乎零成本即可接入现有采样流程。

## 速览

测试时扩展常依赖外部验证器，如编译器或值函数，但在许多实际场景中难以获取高质量验证器。现有基于置信度的无验证器方法在复杂任务上会崩溃，因为高置信度往往意味着探索失败。本文发现稳健的认知搜索需要特定的置信度轨迹模式：初始低置信度以促进探索，最终高置信度以收敛到正确解。据此提出共识性框架，通过组合指标显式评估置信度的时间不对称性，在研究生级数学题和自由形式代码生成任务上超越现有基线。

## 关键点

- 揭示现有置信度型无验证器测试时扩展方法在复杂任务上的关键缺陷：高置信度常对应错误答案。
- 提出置信度轨迹模式：初始低、最终高的时间不对称性更有利于稳健推理。
- 设计共识性框架，组合指标惩罚高初始置信度并要求最终确定性，在数学与代码任务上超越基线。

## 全文

Test-time scaling often uses an external verifier, such as compilers and test cases in coding or trained value functions in robotics applications, to obtain high-quality rollouts. Verifier-free test-time scaling (or VF-TTS) is gaining extensive attention as a mechanism to enhance Large Language Model (LLM) reasoning, primarily because we do not have access to such high-quality verifiers in many real-world applications. Among existing VF-TTS methods, confidence-based VF-TTS methods, which compute and rank rollouts solely by confidence, are particularly promising. Such methods introduce near-zero overhead for sample evaluation and require minimal access to internal model states, making the methods highly flexible across models and tasks. In this paper, we demonstrate a critical limitation of existing confidence-based VF-TTS methods by showing that such methods catastrophically break down on complex tasks. We observe a very interesting phenomenon: uniformly high confidence frequently indicates a failure to explore, favoring confidently wrong answers. To address this, our core insight is that robust cognitive search requires a specific confidence trajectory pattern: such methods perform exploratory branching at the beginning, as manifested by low initial confidence, and converge to a high final confidence solution. To implement this insight, we introduce consilience, a novel selection framework that explicitly evaluates the temporal asymmetry of confidence in reasoning. We operationalize this via a combinatorial metric that actively penalizes high initial confidence while strictly demanding final certainty. Extensive experiments covering both graduate-level mathematics problems and free-form code generation demonstrate that consilience effectively outperforms existing baselines, validating our novel perspective on completion confidence.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_a874f8de9a8bec07b6a8147d