Saved in:
Bibliographic Details
Main Authors: Zhang, Christine, Jurafsky, Dan, Shani, Chen
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.29123
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911715090759680
author Zhang, Christine
Jurafsky, Dan
Shani, Chen
author_facet Zhang, Christine
Jurafsky, Dan
Shani, Chen
contents The next-token prediction (NTP) objective trains language models to predict a single token at each step, even though many continuations can express the same meaning. For example, in the sentence ``this sticker can be placed here'', positioned, attached, or put are all plausible alternatives. While standard NTP training treats these alternatives as mutually exclusive targets, we explore a self-supervised framework that encourages models to predict concepts, approximated as sets of semantically equivalent tokens. Models trained with this concept supervision align better with human similarity judgments, improve classification, clustering, and reranking performance, and achieve comparable or stronger downstream reasoning. These gains come with lower perplexity on semantically meaningful words (Section 3.2) and only minimal increases in global perplexity, suggesting that concepts enhance semantic alignment while preserving language modeling quality. Our code is available at https://anonymous.4open.science/r/learning-concepts-9025 .
format Preprint
id arxiv_https___arxiv_org_abs_2603_29123
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Learning Concepts, Not Tokens: Self-Supervised Semantic Alignment for Language Models
Zhang, Christine
Jurafsky, Dan
Shani, Chen
Computation and Language
The next-token prediction (NTP) objective trains language models to predict a single token at each step, even though many continuations can express the same meaning. For example, in the sentence ``this sticker can be placed here'', positioned, attached, or put are all plausible alternatives. While standard NTP training treats these alternatives as mutually exclusive targets, we explore a self-supervised framework that encourages models to predict concepts, approximated as sets of semantically equivalent tokens. Models trained with this concept supervision align better with human similarity judgments, improve classification, clustering, and reranking performance, and achieve comparable or stronger downstream reasoning. These gains come with lower perplexity on semantically meaningful words (Section 3.2) and only minimal increases in global perplexity, suggesting that concepts enhance semantic alignment while preserving language modeling quality. Our code is available at https://anonymous.4open.science/r/learning-concepts-9025 .
title Learning Concepts, Not Tokens: Self-Supervised Semantic Alignment for Language Models
topic Computation and Language
url https://arxiv.org/abs/2603.29123