arXiv cs.CL· Nelson Jo·· 12 小时前AI 评分31
LLM 的分布对齐及其泛化特性研究
Distributional Alignment of LLMs and Its Generalization Properties
AI 导读
研究提出让 LLM 输出分布与人类参考分布对齐的方法，并在两个大规模基准上验证：模型能实现分布对齐，且可泛化到同领域的未见问题。跨主题泛化取决于主题间关联（如政治问题与机构信任相关），但对齐后的 LLM 可能以出人意料的方式退化，例如丧失创造力，说明其泛化特性本质上不可预测且不稳定。
正文 · 原文
Abstract:Large language models are increasingly used to simulate and probe human behavior, e.g., for surveys. For these applications, a key property is that outputs satisfy distributional alignment: to match its output distribution with some reference (in this case, human) distribution. An obstacle in achieving distributional alignment is generalizing from small training data. In this work, we study (1) how to distributionally align LLMs; and (2) how aligned LLMs might generalize to unseen samples and unrelated topics and tasks. Using two large-scale benchmarks, we find that LLMs can indeed achieve distributional alignment and generalize to unseen questions of the same domain. Cross-topic generalization depends on the relationship between topics (e.g., questions about politics are related to trust in institutions). However, aligned LLMs can degrade in surprising ways (e.g., lose their creative ability), suggesting that its generalization properties are inherently unpredictable and unstable.
| Comments: | IJCAI 2026 Workshop on Augmented Democracies |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.31687 [cs.CL] |
| (or arXiv:2610.31687v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.31687 arXiv-issued DOI via DataCite |
Submission history
From: Nelson Jo [view email]
[v1]
Mon, 5 Oct 2026 12:13:18 UTC (646 KB)
来源：arXiv cs.CL · arxiv.org