---
title: "On Turn-Level Best-of-N Sampling for Scaling LLM Agents"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2611.06868v1"
permalink: "https://feed.astralor.com/items/itm_7f95225d14249d7a37827c6c"
published: "2026-11-04T18:23:14.000Z"
score: 57
channel: "tech"
tags: ["LLM","AI Agent","推理效率","Turn-Level Best-of-N Sampling","Scaling LLM Agents","LLMs","Different"]
---

# On Turn-Level Best-of-N Sampling for Scaling LLM Agents

## 速览

In this work, we evaluate a simple inference-time scaling approach based on \textit{turn-level best-of-N sampling} (\method) for large language model (LLM) agents on complex, long-horizon tasks. We instantiate this usin…

## 关键点

- In this work, we evaluate a simple inference-time scaling approach based on \textit{turn-level best-of-N sampling} (\method) for large language model (LLM) agents on complex, long-horizon tasks.
- We instantiate this using a \textit{generator-selector} framework with two LLMs; at each turn, a \emph{generator} model generates $N$ candidate actions and a \emph{selector} model selects one of them, which is then exec…
- This paper is closely related to the work of \cite{zainullina2025guided} who used critic models for guided search at inference time.

## 全文

In this work, we evaluate a simple inference-time scaling approach based on \textit{turn-level best-of-N sampling} (\method) for large language model (LLM) agents on complex, long-horizon tasks. We instantiate this using a \textit{generator-selector} framework with two LLMs; at each turn, a \emph{generator} model generates $N$ candidate actions and a \emph{selector} model selects one of them, which is then executed in the environment. This paper is closely related to the work of \cite{zainullina2025guided} who used critic models for guided search at inference time. Different from their work, we evaluate if off-the-shelf, open-weight LLMs can be effective at comparing different candidate actions directly. Experiments on long-horizon terminal tasks show that \method can lead to significant performance improvements, both in terms of accuracy (5--10\%) and consistency across repeated attempts. Analysis of agent trajectories reveal a shift in agent behaviour towards desirable patterns such as reduction in number of attempts to solve the task, improved task instruction following, debugging and testing.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_7f95225d14249d7a37827c6c