How Language Models Process Negation

• URL: http://arxiv.org/abs/2605.03052v1
• Date: Mon, 04 May 2026 18:17:05 GMT
• Title: How Language Models Process Negation
• Abstract summary: We study how Large Language Models (LLMs) process negation mechanistically.<n>Our work deepens the understanding of LLMs' internals, highlighting construction-dominant computations and the coexistence of competing mechanisms.
• Score: 56.41361273288279
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We study how Large Language Models (LLMs) process negation mechanistically. First, we establish that even though open-weight models often provide wrong answers to questions involving negation, they do possess internal components that process negation correctly. Their poor accuracy is due to late-layer attention behavior that promotes simple shortcuts; ablating those attention modules greatly improves accuracy on negation-related questions. Second, we uncover how models process negation. We consider two hypotheses: models could use attention heads that attend to the phrase being negated and suppress related concepts, or they could directly construct a representation of the entire negative phrase (e.g., representing "not gas" as a vector that promotes liquids and solids). We apply a range of observational and causal interpretability techniques on Mistral-7B and Llama-3.1-8B to show that models implement both mechanisms, with the "constructive" mechanism being more prominent. Combined, our work deepens the understanding of LLMs' internals, highlighting construction-dominant computations and the coexistence of competing mechanisms within LLMs.

Related papers

• Revisiting the Systematicity in Negation in the Era of In-Context Learning [16.384578665588936]
  We analyze systematicity regarding understanding of negation from two perspectives: behavioral systematicity and representational systematicity.<n>For behavioral systematicity, we confirm that through demonstrations and in-context learning, LLMs can recognize negation expressions and scope within sentences to some extent, but they fail to achieve perfect performance.<n>For representational systematicity, we analyze the extent to which function vectors can be robustly constructed from in-context examples for tasks that are essential to understanding negation.
  arXiv  Detail & Related papers  (2026-06-15T15:44:15Z)
• Mastering Negation: Boosting Grounding Models via Grouped Opposition-Based Learning [87.15765427638195]
  We introduce D-Negation, a new dataset that provides objects annotated with both positive and negative semantic descriptions.<n>We propose a grouped opposition-based learning framework that learns negation-aware representations from limited samples.<n>By fine-tuning fewer than 10 percent of the model parameters, our approach achieves improvements of up to 4.4 mAP and 5.7 mAP on positive and negative semantic evaluations.
  arXiv  Detail & Related papers  (2026-03-13T03:21:48Z)
• Fluid Representations in Reasoning Models [91.77876704697779]
  We present a mechanistic analysis of how QwQ-32B processes abstract structural information.<n>We find that QwQ-32B gradually improves its internal representation of actions and concepts during reasoning.
  arXiv  Detail & Related papers  (2026-02-04T18:34:50Z)
• Training Language Models to Explain Their Own Computations [73.8562596518326]
  We study the extent to which LMs' privileged access to their own internals can be leveraged to produce new techniques for explaining their behavior.<n>Using existing interpretability techniques as a source of ground truth, we fine-tune LMs to generate natural language descriptions of (1) the information encoded by LM features, (2) the causal structure of LMs' internal activations, and (3) the influence of specific input tokens on LM outputs.
  arXiv  Detail & Related papers  (2025-11-11T18:57:14Z)
• How and where does CLIP process negation? [2.5600000778964294]
  We build on the existence task from the VALSE benchmark to test models' understanding of negation. We take inspiration from the literature on model interpretability to explain the behaviour of VL models on the understanding of negation.
  arXiv  Detail & Related papers  (2024-07-15T07:20:06Z)
• Brittle Minds, Fixable Activations: Understanding Belief Representations in Language Models [9.318796743761224]
  Despite growing interest in Theory of Mind (ToM) tasks for evaluating language models (LMs), little is known about how LMs internally represent mental states of self and others.<n>We present the first systematic investigation of belief representations in LMs by probing models across different scales, training regimens, and prompts.<n>Our experiments provide evidence that both model size and fine-tuning substantially improve LMs' internal representations of others' beliefs, which are structured - not mere by-products of spurious correlations - yet brittle to prompt variations.
  arXiv  Detail & Related papers  (2024-06-25T12:51:06Z)
• Revisiting subword tokenization: A case study on affixal negation in large language models [57.75279238091522]
  We measure the impact of affixal negation on modern English large language models (LLMs) We conduct experiments using LLMs with different subword tokenization methods. We show that models can, on the whole, reliably recognize the meaning of affixal negation.
  arXiv  Detail & Related papers  (2024-04-03T03:14:27Z)
• This is not a Dataset: A Large Negation Benchmark to Challenge Large Language Models [4.017326849033009]
  We try to clarify the reasons for the sub-optimal performance of large language models understanding negation. We introduce a large semi-automatically generated dataset of circa 400,000 descriptive sentences about commonsense knowledge. We have used our dataset with the largest available open LLMs in a zero-shot approach to grasp their generalization and inference capability.
  arXiv  Detail & Related papers  (2023-10-24T15:38:21Z)
• Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small [68.879023473838]
  We present an explanation for how GPT-2 small performs a natural language task called indirect object identification (IOI) To our knowledge, this investigation is the largest end-to-end attempt at reverse-engineering a natural behavior "in the wild" in a language model.
  arXiv  Detail & Related papers  (2022-11-01T17:08:44Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.