Image de couverture de livre

From Noise to Narrative: Tracing the Origins of Hallucinations in Transformers

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Suresh, Praneet, Stanley, Jack, Joseph, Sonia, Scimeca, Luca, Bzdok, Danilo
Format:	Preprint
Publié:	2025
Sujets:	Machine Learning Artificial Intelligence
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Quantifying LLM Attention-Head Stability: Implications for Circuit Universality
par: Bali, Karan, et autres
Publié: (2026)

Steering CLIP's vision transformer with sparse autoencoders
par: Joseph, Sonia, et autres
Publié: (2025)

Prisma: An Open Source Toolkit for Mechanistic Interpretability in Vision and Video
par: Joseph, Sonia, et autres
Publié: (2025)

Shaping Inductive Bias in Diffusion Models through Frequency-Based Noise Control
par: Jiralerspong, Thomas, et autres
Publié: (2025)

Learning What Matters: Steering Diffusion via Spectrally Anisotropic Forward Noise
par: Scimeca, Luca, et autres
Publié: (2025)

The Uncanny Valley: A Comprehensive Analysis of Diffusion Models
par: Ghanem, Karam, et autres
Publié: (2024)

Scalable Ensemble Diversification for OOD Generalization and Detection
par: Rubinstein, Alexander, et autres
Publié: (2024)

NoiseFormer -- Noise Diffused Symmetric Attention Transformer
par: Kumar, Phani, et autres
Publié: (2026)

OT-Transformer: A Continuous-time Transformer Architecture with Optimal Transport Regularization
par: Kan, Kelvin, et autres
Publié: (2025)

Mitigating Shortcut Learning with Diffusion Counterfactuals and Diverse Ensembles
par: Scimeca, Luca, et autres
Publié: (2023)

Noise Stability of Transformer Models
par: Haris, Themistoklis, et autres
Publié: (2026)

Estimating Unknown Population Sizes Using the Hypergeometric Distribution
par: Hodgson, Liam, et autres
Publié: (2024)

Transcoders Trace Visual Grounding and Hallucinations in Vision-Language Models
par: Damianos, Dimitrios, et autres
Publié: (2026)

Exploring Loss Design Techniques For Decision Tree Robustness To Label Noise
par: Sztukiewicz, Lukasz, et autres
Publié: (2024)

The Map of Misbelief: Tracing Intrinsic and Extrinsic Hallucinations Through Attention Patterns
par: Hajji, Elyes, et autres
Publié: (2025)

Sample Transform Cost-Based Training-Free Hallucination Detector for Large Language Models
par: Ding, Zeyang, et autres
Publié: (2026)

Uncovering Graph Reasoning in Decoder-only Transformers with Circuit Tracing
par: Dai, Xinnan, et autres
Publié: (2025)

Mechanistic Data Attribution: Tracing the Training Origins of Interpretable LLM Units
par: Chen, Jianhui, et autres
Publié: (2026)

From Noise to Control: Parameterized Diffusion Policies
par: Zhang, Renhao, et autres
Publié: (2026)

From Dispersion to Attraction: Spectral Dynamics of Hallucination Across Whisper Model Scales
par: Viakhirev, Ivan, et autres
Publié: (2026)

Circuit Complexity of Hierarchical Knowledge Tracing and Implications for Log-Precision Transformers
par: Liu, Naiming, et autres
Publié: (2026)

Precision Tracked Transformer via Kalman Filtering, Kriging and Process Noise
par: Long, Bo, et autres
Publié: (2026)

Surgical Knowledge Rewrite in Compact LLMs: An 'Unlearn-then-Learn' Strategy with ($IA^3$) for Localized Factual Modulation and Catastrophic Forgetting Mitigation
par: Ngugi, Stanley
Publié: (2025)

From Flat Facts to Sharp Hallucinations: Detecting Stubborn Errors via Gradient Sensitivity
par: Liew, Yee Zhing, et autres
Publié: (2026)

Priming: Hybrid State Space Models From Pre-trained Transformers
par: Chattopadhyay, Aditya, et autres
Publié: (2026)

ThinkTank: A Framework for Generalizing Domain-Specific AI Agent Systems into Universal Collaborative Intelligence Platforms
par: Surabhi, Praneet Sai Madhu, et autres
Publié: (2025)

Noise-corrected GRPO: From Noisy Rewards to Unbiased Gradients
par: Mansouri, Omar El, et autres
Publié: (2025)

Navigating Shortcuts, Spurious Correlations, and Confounders: From Origins via Detection to Mitigation
par: Steinmann, David, et autres
Publié: (2024)

From Simulation to Enaction: Post-trained language models recognize and react to their own generations
par: G., Asvin, et autres
Publié: (2026)

Operational Wind Speed Forecasts for Chile's Electric Power Sector Using a Hybrid ML Model
par: Suri, Dhruv, et autres
Publié: (2024)

STLGT: A Scalable Trace-Based Linear Graph Transformer for Tail Latency Prediction in Microservices
par: Ding, Yongliang, et autres
Publié: (2026)

Versatile Cardiovascular Signal Generation with a Unified Diffusion Transformer
par: Chen, Zehua, et autres
Publié: (2025)

Cisco Time Series Model Technical Report
par: Gou, Liang, et autres
Publié: (2025)

Hallucination as Trajectory Commitment: Causal Evidence for Asymmetric Attractor Dynamics in Transformer Generation
par: Akarlar, G. Aytug
Publié: (2026)

Prediction of Bank Credit Ratings using Heterogeneous Topological Graph Neural Networks
par: Liu, Junyi, et autres
Publié: (2025)

From Noise to Precision: A Diffusion-Driven Approach to Zero-Inflated Precipitation Prediction
par: Gao, Wentao, et autres
Publié: (2025)

Learning to Reason for Hallucination Span Detection
par: Su, Hsuan, et autres
Publié: (2025)

Grokking as Structural Inference: Transformers Need Bayesian Lottery Tickets
par: Hidajat, Kai, et autres
Publié: (2026)

Origin Tracer: A Method for Detecting LoRA Fine-Tuning Origins in LLMs
par: Liang, Hongyu, et autres
Publié: (2025)

Rethinking Hallucinations: Correctness, Consistency, and Prompt Multiplicity
par: Ganesh, Prakhar, et autres
Publié: (2026)