# LLM-CAS: Dynamic Neuron Perturbation for Real-Time Hallucination Correction

arXiv: https://arxiv.org/abs/2512.18623 · HTML: https://mlsanity.com/paper/2512.18623/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- Builds on 36 of its 55 references that are in the map.
- First posted on arXiv 21 Dec 2025.

## Builds on (most cited first, 30 of 36)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Deep reinforcement learning from human preferences](https://mlsanity.com/paper/1706.03741/) (2017)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension](https://mlsanity.com/paper/1705.03551/) (2017)
- [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022)
- [Trust Region Policy Optimization](https://mlsanity.com/paper/1502.05477/) (2015)
- [Asynchronous Methods for Deep Reinforcement Learning](https://mlsanity.com/paper/1602.01783/) (2016)
- [BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questions](https://mlsanity.com/paper/1905.10044/) (2019)
- [High-Dimensional Continuous Control Using Generalized Advantage Estimation](https://mlsanity.com/paper/1506.02438/) (2015)
- [Inference-Time Intervention: Eliciting Truthful Answers from a Language Model](https://mlsanity.com/paper/2306.03341/) (2023)
- [Steering Llama 2 via Contrastive Activation Addition](https://mlsanity.com/paper/2312.06681/) (2023)
- [A General Language Assistant as a Laboratory for Alignment](https://mlsanity.com/paper/2112.00861/) (2021)
- [Mass-Editing Memory in a Transformer](https://mlsanity.com/paper/2210.07229/) (2022)
- [Training Diffusion Models with Reinforcement Learning](https://mlsanity.com/paper/2305.13301/) (2023)
- [Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2307.15217/) (2023)
- [Knowledge Neurons in Pretrained Transformers](https://mlsanity.com/paper/2104.08696/) (2021)
- [Fast Model Editing at Scale](https://mlsanity.com/paper/2110.11309/) (2021)
- [A Corpus and Evaluation Framework for Deeper Understanding of Commonsense Stories](https://mlsanity.com/paper/1604.01696/) (2016)
- [Editing Large Language Models: Problems, Methods, and Opportunities](https://mlsanity.com/paper/2305.13172/) (2023)
- [Hierarchical Deep Reinforcement Learning: Integrating Temporal Abstraction and Intrinsic Motivation](https://mlsanity.com/paper/1604.06057/) (2016)
- [Follow-Your-Emoji: Fine-Controllable and Expressive Freestyle Portrait Animation](https://mlsanity.com/paper/2406.01900/) (2024)
- [Sources of Hallucination by Large Language Models on Inference Tasks](https://mlsanity.com/paper/2305.14552/) (2023)
- [Unfamiliar Finetuning Examples Control How Language Models Hallucinate](https://mlsanity.com/paper/2403.05612/) (2024)
- [Semantics-Adaptive Activation Intervention for LLMs via Dynamic Steering Vectors](https://mlsanity.com/paper/2410.12299/) (2024)
- [MagicStick: Controllable Video Editing via Control Handle Transformations](https://mlsanity.com/paper/2312.03047/) (2023)
- [Towards better understanding of gradient-based attribution methods for Deep Neural Networks](https://mlsanity.com/paper/1711.06104/) (2017)

## Built on by (most cited first, 15 of 15)

- [AdaMARP: An Adaptive Multi-Agent Interaction Framework for General Immersive Role-Playing](https://mlsanity.com/paper/2601.11007/) (2026)
- [CoAgent: Collaborative Planning and Consistency Agent for Coherent Video Generation](https://mlsanity.com/paper/2512.22536/) (2025)
- [ResWM: Residual-Action World Model for Visual RL](https://mlsanity.com/paper/2603.11110/) (2026)
- [Neuron-Anchored Rule Extraction for Large Language Models via Contrastive Hierarchical Ablation](https://mlsanity.com/paper/2605.03058/) (2026)
- [MoE-DiffuSeq: Enhancing Long-Document Diffusion Models with Sparse Attention and Mixture of Experts](https://mlsanity.com/paper/2512.20604/) (2025)
- [SirenPose: Dynamic Scene Reconstruction via Geometric Supervision](https://mlsanity.com/paper/2512.20531/) (2025)
- [SA-DiffuSeq: Addressing Computational and Scalability Challenges in Long-Document Generation with Sparse Attention](https://mlsanity.com/paper/2512.20724/) (2025)
- [A Three-Level Alignment Framework for Large-Scale 3D Retrieval and Controlled 4D Generation](https://mlsanity.com/paper/2512.22294/) (2025)
- [Learning Dynamic Scene Reconstruction with Sinusoidal Geometric Priors](https://mlsanity.com/paper/2512.22295/) (2025)
- [Towards Efficient Post-Training via Fourier-Driven Adapter Architectures](https://mlsanity.com/paper/2512.22378/) (2025)
- [Self-Rewarded Multimodal Coherent Reasoning Across Diverse Visual Domains](https://mlsanity.com/paper/2512.22545/) (2025)
- [Multi-Agent Cooperative Learning for Robust Vision-Language Alignment under OOD Concepts](https://mlsanity.com/paper/2601.09746/) (2026)
- [Bridging the Semantic Chasm: Synergistic Conceptual Anchoring for Generalized Few-Shot and Zero-Shot OOD Perception](https://mlsanity.com/paper/2602.00340/) (2026)
- [A Scalable Curiosity-Driven Game-Theoretic Framework for Long-Tail Multi-Label Learning in Data Mining](https://mlsanity.com/paper/2602.15330/) (2026)
- [ProgAgent:A Continual RL Agent with Progress-Aware Rewards](https://mlsanity.com/paper/2603.07784/) (2026)

## Abstract (arXiv)

Large language models (LLMs) often generate hallucinated content that lacks factual or contextual grounding, limiting their reliability in critical applications. Existing approaches such as supervised fine-tuning and reinforcement learning from human feedback are data intensive and computationally expensive, while static parameter editing methods struggle with context dependent errors and catastrophic forgetting. We propose LLM-CAS, a framework that formulates real-time hallucination correction as a hierarchical reinforcement learning problem. LLM-CAS trains an agent to learn a policy that dynamically selects temporary neuron perturbations during inference based on the current context. Unlike prior dynamic approaches that rely on heuristic or predefined adjustments, this policy driven mechanism enables adaptive and fine grained correction without permanent parameter modification. Experiments across multiple language models demonstrate that LLM-CAS consistently improves factual accuracy, achieving gains of 10.98 percentage points on StoryCloze, 2.71 points on TriviaQA, and 2.06 points on the MC1 score of TruthfulQA. These results outperform both static editing methods such as ITI and CAA and the dynamic SADI framework. Overall, LLM-CAS provides an efficient and context aware solution for improving the reliability of LLMs, with promising potential for future multimodal extensions.