# Don't Think of the White Bear: Ironic Negation in Transformer Models Under Cognitive Load

arXiv: https://arxiv.org/abs/2511.12381 · HTML: https://mlsanity.com/paper/2511.12381/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning.
- Builds on 12 of its 24 references that are in the map.
- First posted on arXiv 15 Nov 2025.

## Placed in

- LLM reasoning / [Reasoning benchmarks and evaluation](https://mlsanity.com/lr/reasoning-benchmarks-and-evaluation/)

## Builds on (most cited first, 12 of 12)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [OPT: Open Pre-trained Transformer Language Models](https://mlsanity.com/paper/2205.01068/) (2022)
- [Lost in the Middle: How Language Models Use Long Contexts](https://mlsanity.com/paper/2307.03172/) (2023)
- [BLOOM: A 176B-Parameter Open-Access Multilingual Language Model](https://mlsanity.com/paper/2211.05100/) (2022)
- [gpt-oss-120b & gpt-oss-20b Model Card](https://mlsanity.com/paper/2508.10925/) (2025)
- [Copy Suppression: Comprehensively Understanding an Attention Head](https://mlsanity.com/paper/2310.04625/) (2023)
- [Investigating and Addressing Hallucinations of LLMs in Tasks Involving Negation](https://mlsanity.com/paper/2406.05494/) (2024)
- [Mitigating Copy Bias in In-Context Learning through Neuron Pruning](https://mlsanity.com/paper/2410.01288/) (2024)
- [Negation: A Pink Elephant in the Large Language Models' Room?](https://mlsanity.com/paper/2503.22395/) (2025)
- [Thunder-NUBench: A Benchmark for LLMs' Sentence-Level Negation Understanding](https://mlsanity.com/paper/2506.14397/) (2025)
- [Do not think about pink elephant!](https://mlsanity.com/paper/2404.15154/) (2024)

## Built on by (most cited first, 2 of 2)

- [Can You Keep a Secret? Involuntary Information Leakage in Language Model Writing](https://mlsanity.com/paper/2605.10794/) (2026)
- [From Caricature to Collective Artifact: How Persona Construction Drives Polarization in Generative Social Simulations](https://mlsanity.com/paper/2610.16498/) (2026)

## Abstract (arXiv)

Negation instructions such as 'do not mention X' can paradoxically increase the accessibility of X in human thought, a phenomenon known as ironic rebound. Large language models (LLMs) face the same challenge: suppressing a concept requires internally activating it, which may prime rebound instead of avoidance. We investigated this tension with two experiments. (1) Load & content: after a negation instruction, we vary distractor text (semantic, syntactic, repetition) and measure rebound strength. (2) Polarity separation: We test whether models distinguish neutral from negative framings of the same concept and whether this separation predicts rebound persistence. Results show that rebound consistently arises immediately after negation and intensifies with longer or semantic distractors, while repetition supports suppression. Stronger polarity separation correlates with more persistent rebound. Together, these findings, complemented by a circuit tracing analysis that identifies sparse middle-layer attention heads amplifying forbidden tokens while early layers suppress, link cognitive predictions of ironic rebound with mechanistic insights into long-context interference. To support future work, we release ReboundBench, a dataset of 5,000 systematically varied negation prompts designed to probe rebound in LLMs.