mlsanityData to 7 Oct 2026

Hallucination Basins: A Dynamic Framework for Understanding and Controlling LLM Hallucinations

arXiv 2604.04743 ↗First posted 6 Apr 2026cs.CL cs.AI cs.SYcited by 2

Not on the map. Why?

Abstract (arXiv)

Large language models (LLMs) hallucinate: they produce fluent outputs that are factually incorrect. We present a geometric dynamical systems framework in which hallucinations arise from task-dependent basin structure in latent space. Using autoregressive hidden-state trajectories across multiple open-source models and benchmarks, we find that separability is strongly task-dependent rather than universal: factoid settings can show clearer basin separation, whereas summarization and misconception-heavy settings are typically less stable and often overlap. We formalize this behavior with task-complexity and multi-basin theorems, characterize basin emergence in L-layer transformers, and show that geometry-aware steering can reduce hallucination probability without retraining.

Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 18 of its 34 references that are in the map. First posted on arXiv 6 Apr 2026.

Builds on

18 of 34 references.

References · 18

  1. 2025Qwen3 Technical Reportcited by 8,195
  2. 2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,919
  3. 2024Gemma 2: Improving Open Language Models at a Practical Sizecited by 1,928
  4. 2023A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questionscited by 1,582
  5. 2018FEVER: a large-scale dataset for Fact Extraction and VERificationcited by 1,362
  6. 2021MuSiQue: Multihop Questions via Single-hop Question Compositioncited by 893
  7. 2023SelfCheckGPT: Zero-Resource Black-Box Hallucination Detection for Generative Large Language Modelscited by 787
  8. 2020Hopfield Networks is All You Needcited by 461
  9. 2024INSIDE: LLMs' Internal States Retain the Power of Hallucination Detectioncited by 317
  10. 2024ReDeEP: Detecting Hallucination in Retrieval-Augmented Generation via Mechanistic Interpretabilitycited by 100
  11. 2024Adaptive Activation Steering: A Tuning-Free LLM Truthfulness Improvement Method for Diverse Hallucinations Categoriescited by 74
  12. 2024Look Twice Before You Answer: Memory-Space Visual Retracing for Hallucination Mitigation in Multimodal Large Language Modelscited by 72
6 more
  1. 2025Steer LLM Latents for Hallucination Detectioncited by 49
  2. 2024In-Context Sharpness as Alerts: An Inner Representation Perspective for Hallucination Mitigationcited by 43
  3. 2025Why and How LLMs Hallucinate: Connecting the Dots with Subsequence Associationscited by 14
  4. 2025(Im)possibility of Automated Hallucination Detection in Large Language Modelscited by 14
  5. 2025Memorisation and forgetting in a learning Hopfield neural network: bifurcation mechanisms, attractors and basinscited by 2
  6. 2025Associative Memory Model with Neural Networks: Memorizing multiple images with one neuroncited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Understanding Why Language Models Hallucinate: Testing Reasoning Against Priorscited by 1
  2. 2026Prediction of Prediction (PoP): Inter-Layer Activation Fusion for Single-Pass Hallucination Detection in Large Language Modelscited by 0