Representational Drift

Follow this topic Ask about this topic

Recent momentum

-86%

1 papers in the last 28 days · 0.0% of indexed attention

Twelve weeks of publication activity for this topic as it is defined today.

Weekly history

Recent digests

What was published in this topic, kept on the site without email delivery.

Period ending 2026-09-07

2 new papers

A weekly snapshot of new work published in Representational Drift.

47 papers

Latest in Representational Drift

Open your feed →

How Perturbations Propagate: A Multi-Level Analysis of Robustness in Large Language Models
Sep 3, 2026Dun Li Chan, Emily Liu, Niyathi Allu +1Representational Drift
Representational separation between unitary and channel quantum generative models via shared classical randomness at shallow depth
Aug 5, 2026Arunava Majumder, Marius Krumm, Hendrik Poulsen Nautrup +1Quantum Circuits Generative Models
Isotropy Cliffs: The Geometric Signature of Decision-Making in Large Language Models
Aug 1, 2026Okan S. Coskun, Florian Rottach, Carsten Eickhoff +1Multiple-Choice Questions Decisions
The Learning Objective Governs Perceptual Narrowing: A Cross-Lingual, Layer-Wise, Ten-Seed Study of Self-Supervised Speech Encoders
Aug 1, 2026Sejin YooSelf-Supervised Speech Models Representational Drift
Do Unified Multimodal Models Think in One Space? A Lens Through Cross-Branch Steering
Jul 29, 2026Yu Wang, Sharon LiRepresentational Drift Lens
The Confounder Trap: Treatment-Encoding Representations in Causal Inference with Text
Jul 28, 2026Marie Neubrander, Graham Tierney, Alexander VolfovskyLatent Confounders Causal Inferences
Probing the Origins of Reasoning Performance: Representational Quality for Mathematical Problem-Solving in RL vs. SFT Fine-Tuned Models
Jul 28, 2026Antyabha Rahman, Akshaj Gurugubelli, Omar Ankit +2Large Reasoning Models Supervised Fine-Tuning
What CLIP Knows but Cannot Say: Recovering Negation from Frozen Intermediate Features
Jul 25, 2026Chen-Yi Lu, Yueh-Shao Chen, Somali ChaterjiContrastive Language-Image Pre-Training Negation
Mixture-of-Control: State-Aware Fine-Tuning for Transformer-based Models
Jun 30, 2026Duc Anh Nguyen, Tien Ngoc Luu, Tung Pham +1Parameter-Efficient Fine-Tuning Methods Finetuning
A Gravitational Interpretation of Fine-Tuning Reversion
Jun 26, 2026Samuele Poppi, Nils LukasSafety Alignment Finetuning
What Does a Chemical Language Model Know About Molecules?
Jun 22, 2026Christian Kenneth, Etowah Adams, Liam Bai +1Molecular Structure Elucidation Protein Language Model
Lexical Consensus: Grounded Word Learning and Shared Meaning in Artificial Agents
Jun 20, 2026Patricio M. VeraLexicon Contextual Grounding
The Geometry of Saturation: Effective Rank Predicts When Labels Stop Helping in Few-Shot Classification
Jun 12, 2026Arnav GuptaSaturation Spearman Correlation
The Stable Recovery Manifold: Geometric Principles Governing Recoverability in Continual Learning
Jun 11, 2026Ayushman Trivedi, Bhavika MelwaniCatastrophic Forgetting Continual Learning
Detecting Explanatory Insufficiency in Learned Representations: A Framework for Representational Vigilance
Jun 11, 2026Jacques Raynal, Pierre Slangen, Elsa Raynal +1Representation Learning Representational Drift
When Role-playing, Do Models Believe What They Say?
Jun 9, 2026Benjamin Sturgeon, David Africa, Sid BlackRole-Playing Personas
Bootstrap Theory of Representational Emergence: Explanatory Insufficiency as a Driver of Representation Learning and World Models
Jun 5, 2026Jacques Raynal, Pierre Slangen, Elsa Raynal +1Representational Drift Representational Capacity
Trajectory Dynamics in Language Model Hidden States Predict Human Processing Costs Beyond Surprisal
Jun 3, 2026Elan BarenholtzProtein Language Model Surprisal
Self-Revising Discovery Systems for Science: A Categorical Framework for Agentic Artificial Intelligence
May 31, 2026Fiona Y. Wang, Markus J. BuehlerAgentic Discovery Scientific Discovery
Trait-space Monitoring for Emergent Misalignment During Supervised Finetuning
May 31, 2026Huy Nghiem, Sy-Tuyen Ho, Sarah Wiegreffe +1Misalignment Finetuning
Cognitive Fatigue in Autoregressive Transformers: Formalization and Measurement
May 29, 2026Riju Marwah, Ritvik Garimella, Vishal Pallagani +3Autoregressive Language Models Large Language Models Fail
The Attentional White Bear Effect in Transformer Language Models
May 27, 2026Rebecca Ramnauth, Brian ScassellatiProtein Language Model Representational Drift
Feature Geometry of LoRA Adapters: A Sparse Autoencoder Analysis of Representational Divergence in Fine-Tuned Language Models
May 27, 2026Prasanth K KSparse Autoencoders Model Activations
Real Images, Worse Judgments: Evaluating Vision-Language Models on Concreteness and Imagery
May 26, 2026Yifan Jiang, Ruoxi Ning, Sheng Yao +1Representational Drift Relevance
RepSAM: Bridging Foundation Models to Robotic Vision via Representation-Guided Adaptation
May 25, 2026Wenhui ChuRobot Perception Recent Vision Foundation Models
Convergence Without Understanding: When Language Models Agree on Representations but Disagree on Reasoning
May 22, 2026Muhammad Usama, Dong Eui ChangRepresentational Drift Optimization Convergence Analysis
Feature Lottery? A Bifurcation Theory of Concept Emergence
May 22, 2026Fuming YangFeature Learning Bifurcation
TextReg: Mitigating Prompt Distributional Overfitting via Regularized Text-Space Optimization
May 20, 2026Lucheng Fu, Ye Yu, Yiyang Wang +4Overfitting Representational Drift
Brain alignment of reasoning and action representations from vision-language and action models during naturalistic gameplay
May 19, 2026Subba Reddy Oota, Anant Khandelwal, Khushbu Pahwa +4Human Visual Cortex Representational Drift
OmniISR: A Unified Framework for Centralized and Federated Learning via Intermediate Supervision and Regularization
May 19, 2026Wei-Bin Kou, Guangxu Zhu, Ming Tang +4Semi-Supervised Learning Regularization Methods
Mechanistic Interpretability of EEG Foundation Models via Sparse Autoencoders
May 13, 2026William Lehn-Schiøler, Magnus Ruud Kjær, Rahul Thapa +10Electroencephalography Foundation Models Improving Sparse Autoencoders
Towards Visually Grounded Multimodal Summarization via Cross-Modal Transformer and Gated Attention
May 12, 2026Abid Ali, Diego Molla-Aliod, Usman NaseemSummarization Representational Drift
A Mechanistic Investigation of Supervised Fine Tuning
May 12, 2026Ruhaan ChopraSupervised Fine-Tuning Model Activations
Lost or Hidden? A Concept-Level Forgetting in Supervised Continual Learning
May 10, 2026Katarzyna Filus, Kamil Faber, Roberto Corizzo +1Forgetting Continual Learning
XTinyU-Net: Training-Free U-Net Scaling via Initialization-Time Sensitivity
May 10, 2026Alvin Kimbowa, Moein Heidari, David Liu +1U-Net Convolutional Neural Network Compression
Mechanistic Analysis of Alignment Algorithms in Language Models
May 9, 2026Aarush Sinha, Ishan Garg, Veeraraju Elluru +2Large Language Model Alignment Representational Drift
Formalizing Latent Thoughts: Four Axioms of Thought Representation in LLMs
May 7, 2026Fahd Seddik, Fatemeh FardLatent Thoughts Representational Drift
AHPA: Adaptive Hierarchical Prior Alignment for Diffusion Transformers
May 5, 2026Ruibin Min, Yexin Liu, Aimin Pan +5Diffusion Transformers Denoising Trajectory
Lost in State Space: Probing Frozen Mamba Representations
Apr 30, 2026Bhagyashree Wagh, Akash SinghHidden States Recurrent State
When 2D Tasks Meet 1D Serialization: On Serialization Friction in Structured Tasks
Apr 29, 2026Chung-Hsiang Lo, Lu Li, Diji Yang +4Representational Drift Large Output Spaces
Monitoring Neural Training with Topology: A Footprint-Predictable Collapse Index
Apr 28, 2026Alexander KalinowskiNeural Collapse Topology
Why Are Some Emotions Harder for LLMs? Uncovering the Causal Mechanisms of Emotion Inference via Sparse Autoencoders
Apr 28, 2026Bangzhao Shu, Arinjay Singh, Mai ElSheriefEmotion Large Language Models Fail
Rank, Head-Channel Non-Identifiability, and Symmetry Breaking: A Precise Analysis of Representational Collapse in Transformers
Apr 26, 2026Giansalvo CirrincioneMulti-Head Attention Batch Normalization
To See the Unseen: on the Generalization Ability of Transformers in Symbolic Reasoning
Apr 23, 2026Nevena Lazić, Liam Fowl, András György +1Transformer Encoder Vector Embeddings
REZE: Representation Regularization for Domain-adaptive Text Embedding Pre-finetuning
Apr 19, 2026Seungmin Lee, Jeonghwan Lee, Hyunkuk Lim +2Patient-Specific Textual Embeddings Domain Adaptation
Decomposing the Depth Profile of Fine-Tuning
Apr 19, 2026Jayadev BillaFinetuning Neural Architectures
Can Fine-Tuning Erase Your Edits? On the Fragile Coexistence of Knowledge Editing and Adaptation
Nov 8, 2025Yinjie Cheng, Paul Youssef, Christin Seifert +2Multimodal Knowledge Editing Global Editing

Sep 3, 2026 · cs.CLJ/K move · Enter open · S save

How Perturbations Propagate: A Multi-Level Analysis of Robustness in Large Language Models

Dun Li Chan, Emily Liu, Niyathi Allu, Christian Hoang

Language models encounter typos, corrupted text, altered words, and disrupted token order, yet robustness is usually evaluated only through output behavior. We study how six naturalistic and synthetic input perturbations propagate through decoder-only language models at three levels: output behavior, hidden-state geometry, and attention-head function. We evaluate behavioral effects across four GPT-2 and two Qwen2.5 checkpoints by analyzing layerwise geometry using centered kernel alignment and intrinsic dimension, and examine attention-head responses in GPT-2. Perturbation types produce distinguishable metric profiles that are not fully captured by output measures and are only partly consistent across the tested checkpoints. Copying scores are especially associated with activation-patching recovery under token substitution and shuffling. Gradient-guided HotFlip perturbations also cause stronger behavioral and representational disruption than rate-matched random token substitutions in GPT-2; their behavioral effects are consistent across all six tested checkpoints. Our results show that robustness claims based on a single behavioral or representational metric can be misleading, and motivate multi-level evaluation of how perturbations alter language-model computation.