cs.CLOct 15, 2021

Coherence boosting: When your pretrained language model is not paying enough attention

Authors: Esmeralda S. Whitammer, Zhen Wang, Nebojsa Jojic

Organizations: Ohio State University · Microsoft Research

Abstract

Long-range semantic coherence remains a challenge in automatic language generation and understanding. We demonstrate that large language models have insufficiently learned the effect of distant words on next-token prediction. We present coherence boosting, an inference procedure that increases a LM's focus on a long context. We show the benefits of coherence boosting with pretrained models by distributional analyses of generated ordinary text and dialog responses. It is also found that coherence boosting with state-of-the-art models for various zero-shot NLP tasks yields performance gains with no additional training.

Explore similar work

CardsList
  1. Does Linguistic Structure Enrichment Enhance Coherence Assessment? Not With Current Architectures

    Sep 9, 2026Victor Mazzotti, Luiz Pereira, Marina Bitencourt dos Santos +4LLM EvaluationFake News Detection

  2. Self-Improvement as Coherence Optimization: A Theoretical Account

    Jan 20, 2026Tianyi Qiu, Ahmed Hani Ismail, Zhonghao He +1Language Model Self-Improvement

  3. IS-CoT: Breaking the Long-form Generation Collapse via Interleaved Structural Thinking

    Jun 8, 2026Zechen Sun, Yuyang Sun, Zecheng Tang +6Open-Ended GenerationLong-Form Document Generation