# LLM 多轮信息寻求评估研究

- 来源：Rohan Paul (@rohanpaul_ai)
- 发布时间：2026-09-13 10:48
- AIHOT 分数：40
- AIHOT 链接：https://aihot.news/items/cmtz8nh190yqiroupi69fb0li
- 原文链接：https://x.com/rohanpaul_ai/status/2098966898742349888

## AI 摘要

LLM 经常低估自己需要多少信息，所以要测试它们何时停止，而不是仅依赖答案准确率。

正确答案可能掩盖信息收集的不足。

模型可能利用先验知识或猜对答案，而没有收集足够的证据。

– arxiv. org/abs/2608.14808

标题："Do LLMs Know What to Ask and When? Evaluating Multi-Turn Information Seeking"

## 正文

LLMs frequently underestimate how much information they need, so test when they stop instead of relying on answer accuracy alone.

A correct answer can hide poor information gathering.

The model may use prior knowledge or guess correctly without collecting enough evidence.

– arxiv. org/abs/2608.14808

Title: "Do LLMs Know What to Ask and When? Evaluating Multi-Turn Information Seeking"