Image de couverture de livre

The Geometric Canary: Predicting Steerability and Detecting Drift via Representational Stability

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteur principal:	Raju, Prashant C.
Format:	Preprint
Publié:	2026
Sujets:	Machine Learning Computation and Language
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Geometric Stability: The Missing Axis of Representations
par: Raju, Prashant C.
Publié: (2026)

Exploring Chain-of-Thought Reasoning for Steerable Pluralistic Alignment
par: Zhang, Yunfan, et autres
Publié: (2025)

Bayesian Preference Learning for Test-Time Steerable Reward Models
par: Hong, Jiwoo, et autres
Publié: (2026)

A Course Correction in Steerability Evaluation: Revealing Miscalibration and Side Effects in LLMs
par: Chang, Trenton, et autres
Publié: (2025)

Integrating Time Series into LLMs via Multi-layer Steerable Embedding Fusion for Enhanced Forecasting
par: Chen, Zhuomin, et autres
Publié: (2025)

Inner Speech as Behavior Guides: Steerable Imitation of Diverse Behaviors for Human-AI coordination
par: Trivedi, Rakshit, et autres
Publié: (2026)

PRISM: A Geometric Risk Bound that Decomposes Drift into Scale, Shape, and Head
par: Lin, Chieh-Yen, et autres
Publié: (2026)

The Geometric Alignment Tax: Tokenization vs. Continuous Geometry in Scientific Foundation Models
par: Raju, Prashant C.
Publié: (2026)

Conditional Language Policy: A General Framework for Steerable Multi-Objective Finetuning
par: Wang, Kaiwen, et autres
Publié: (2024)

Event2Vec: A Geometric Approach to Learning Composable Representations of Event Sequences
par: Sulc, Antonin
Publié: (2025)

Tool Calling is Linearly Readable and Steerable in Language Models
par: Wu, Zekun, et autres
Publié: (2026)

Understanding Knowledge Drift in LLMs through Misinformation
par: Fastowski, Alina, et autres
Publié: (2024)

The Geometric Price of Discrete Logic: Context-driven Manifold Dynamics of Number Representations
par: Zhang, Long, et autres
Publié: (2026)

Breaking Quadratic Barriers: A Non-Attention LLM for Ultra-Long Context Horizons
par: Kiruluta, Andrew, et autres
Publié: (2025)

XLGoBench: Detecting cross-lingual skill gaps with algorithmic tasks
par: Jain, Purvam, et autres
Publié: (2026)

S2 Chunking: A Hybrid Framework for Document Segmentation Through Integrated Spatial and Semantic Analysis
par: Verma, Prashant
Publié: (2025)

Drifting Objectives for Refining Discrete Diffusion Language Models
par: Oba, Daisuke, et autres
Publié: (2026)

Inductive Entity Representations from Text via Link Prediction
par: Daza, Daniel, et autres
Publié: (2020)

The Canary's Echo: Auditing Privacy Risks of LLM-Generated Synthetic Text
par: Meeus, Matthieu, et autres
Publié: (2025)

The Confidence Manifold: Geometric Structure of Correctness Representations in Language Models
par: Cho, Seonglae, et autres
Publié: (2026)

Drift-to-Action Controllers: Budgeted Interventions with Online Risk Certificates
par: Lamaakal, Ismail, et autres
Publié: (2026)

InnerThoughts: Disentangling Representations and Predictions in Large Language Models
par: Chételat, Didier, et autres
Publié: (2025)

Joint Detection of Fraud and Concept Drift inOnline Conversations with LLM-Assisted Judgment
par: Senol, Ali, et autres
Publié: (2025)

Scale Determines Whether Language Models Organize Representation Geometry for Prediction
par: Xu, Weilun
Publié: (2026)

Foundation Model's Embedded Representations May Detect Distribution Shift
par: Vargas, Max, et autres
Publié: (2023)

Cross-Lingual Prompt Steerability: Towards Accurate and Robust LLM Behavior across Languages
par: Zhang, Lechen, et autres
Publié: (2025)

Protect$^*$: Steerable Retrosynthesis through Neuro-Symbolic State Encoding
par: Sathyanarayana, Shreyas Vinaya, et autres
Publié: (2026)

Stabilizing Policy Optimization via Logits Convexity
par: Chen, Hongzhan, et autres
Publié: (2026)

Silence the Judge: Reinforcement Learning with Self-Verifier via Latent Geometric Clustering
par: Zhang, Nonghai, et autres
Publié: (2026)

A Lightweight Method to Disrupt Memorized Sequences in LLM
par: Prashant, Parjanya Prajakta, et autres
Publié: (2025)

Patent Representation Learning via Self-supervision
par: Zuo, You, et autres
Publié: (2025)

Fair Text Classification via Transferable Representations
par: Leteno, Thibaud, et autres
Publié: (2025)

Objective Matters: Fine-Tuning Objectives Shape Safety, Robustness, and Persona Drift
par: Vennemeyer, Daniel, et autres
Publié: (2026)

Few-Shot Detection of Machine-Generated Text using Style Representations
par: Soto, Rafael Rivera, et autres
Publié: (2024)

Demystifying When Pruning Works via Representation Hierarchies
par: He, Shwai, et autres
Publié: (2026)

I Can't Believe It's Not Robust: Catastrophic Collapse of Safety Classifiers under Embedding Drift
par: Sahoo, Subramanyam, et autres
Publié: (2026)

Implicit Geometry of Next-token Prediction: From Language Sparsity Patterns to Model Representations
par: Zhao, Yize, et autres
Publié: (2024)

Mitigating Semantic Drift: Evaluating LLMs' Efficacy in Psychotherapy through MI Dialogue Summarization
par: Kumar, Vivek, et autres
Publié: (2025)

Model-Agnostic Sentiment Distribution Stability Analysis for Robust LLM-Generated Texts Detection
par: Li, Siyuan, et autres
Publié: (2025)

Methods for Generating Drift in Text Streams
par: Garcia, Cristiano Mesquita, et autres
Publié: (2024)