Copertina

Linear representations in language models can change dramatically over a conversation

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Lampinen, Andrew Kyle, Li, Yuxuan, Hosseini, Eghbal, Bhardwaj, Sangnie, Shanahan, Murray
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Computation and Language Machine Learning
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

The broader spectrum of in-context learning
di: Lampinen, Andrew Kyle, et al.
Pubblicazione: (2024)

Context Structure Reshapes the Representational Geometry of Language Models
di: Hosseini, Eghbal A., et al.
Pubblicazione: (2026)

Just-in-time and distributed task representations in language models
di: Li, Yuxuan, et al.
Pubblicazione: (2025)

On the generalization of language models from in-context learning and finetuning: a controlled study
di: Lampinen, Andrew K., et al.
Pubblicazione: (2025)

Latent learning: episodic memory complements parametric learning by enabling flexible reuse of experiences
di: Lampinen, Andrew Kyle, et al.
Pubblicazione: (2025)

How do language models learn facts? Dynamics, curricula and hallucinations
di: Zucchet, Nicolas, et al.
Pubblicazione: (2025)

The in-context inductive biases of vision-language models differ across modalities
di: Allen, Kelsey, et al.
Pubblicazione: (2025)

Still "Talking About Large Language Models": Some Clarifications
di: Shanahan, Murray
Pubblicazione: (2024)

Representational Curvature Modulates Behavioral Uncertainty in Large Language Models
di: King, Jack, et al.
Pubblicazione: (2026)

Beneath the Surface: Investigating LLMs' Capabilities for Communicating with Subtext
di: Ahuja, Kabir, et al.
Pubblicazione: (2026)

Interpretability Illusions in the Generalization of Simplified Models
di: Friedman, Dan, et al.
Pubblicazione: (2023)

Learned feature representations are biased by complexity, learning order, position, and more
di: Lampinen, Andrew Kyle, et al.
Pubblicazione: (2024)

Realised Volatility Forecasting: Machine Learning via Financial Word Embedding
di: Rahimikia, Eghbal, et al.
Pubblicazione: (2021)

Representation biases: will we achieve complete understanding by analyzing representations?
di: Lampinen, Andrew Kyle, et al.
Pubblicazione: (2025)

The representation landscape of few-shot learning and fine-tuning in large language models
di: Doimo, Diego, et al.
Pubblicazione: (2024)

Inducing anxiety in large language models can induce bias
di: Coda-Forno, Julian, et al.
Pubblicazione: (2023)

Do different prompting methods yield a common task representation in language models?
di: Davidson, Guy, et al.
Pubblicazione: (2025)

Large language models reorganize representational geometry during in-context learning
di: Xiong, Hua-Dong, et al.
Pubblicazione: (2026)

State space models can express n-gram languages
di: Nandakumar, Vinoth, et al.
Pubblicazione: (2023)

Distinct Computations Emerge From Compositional Curricula in In-Context Learning
di: Lee, Jin Hwa, et al.
Pubblicazione: (2025)

Comparison of different Unique hard attention transformer models by the formal languages they can recognize
di: Ryvkin, Leonid
Pubblicazione: (2025)

Perturbation: A simple and efficient adversarial tracer for representation learning in language models
di: Rozner, Joshua, et al.
Pubblicazione: (2026)

Symmetry in language statistics shapes the geometry of model representations
di: Karkada, Dhruva, et al.
Pubblicazione: (2026)

Modeling the language cortex with form-independent and enriched representations of sentence meaning reveals remarkable semantic abstractness
di: Saha, Shreya, et al.
Pubblicazione: (2025)

Large language models can accurately predict searcher preferences
di: Thomas, Paul, et al.
Pubblicazione: (2023)

Large language models can learn and generalize steganographic chain-of-thought under process supervision
di: Skaf, Joey, et al.
Pubblicazione: (2025)

Transformers need glasses! Information over-squashing in language tasks
di: Barbero, Federico, et al.
Pubblicazione: (2024)

Aligning language models with human preferences
di: Korbak, Tomasz
Pubblicazione: (2024)

Evaluating language models as risk scores
di: Cruz, André F., et al.
Pubblicazione: (2024)

Benchmarked Yet Not Measured -- Generative AI Should be Evaluated Against Real-World Utility
di: Mondal, Ishani, et al.
Pubblicazione: (2026)

Language models show human-like content effects on reasoning tasks
di: Dasgupta, Ishita, et al.
Pubblicazione: (2022)

Amortizing intractable inference in large language models
di: Hu, Edward J., et al.
Pubblicazione: (2023)

Linear probes rely on textual evidence: Results from leakage mitigation studies in language models
di: Boxo, Gerard, et al.
Pubblicazione: (2025)

EEG-CLIP : Learning EEG representations from natural language descriptions
di: Ndir, Tidiane Camaret, et al.
Pubblicazione: (2025)

Anchor function: a type of benchmark functions for studying language models
di: Zhang, Zhongwang, et al.
Pubblicazione: (2024)

Parallelizing Linear Transformers with the Delta Rule over Sequence Length
di: Yang, Songlin, et al.
Pubblicazione: (2024)

A framework for analyzing concept representations in neural models
di: Naowarat, Burin, et al.
Pubblicazione: (2026)

Perturbed examples reveal invariances shared by language models
di: Rawal, Ruchit, et al.
Pubblicazione: (2023)

A mean teacher algorithm for unlearning of language models
di: Klochkov, Yegor
Pubblicazione: (2025)

Do language models plan ahead for future tokens?
di: Wu, Wilson, et al.
Pubblicazione: (2024)