Cover Image

Fresh in memory: Training-order recency is linearly encoded in language model activations

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Krasheninnikov, Dmitrii, Turner, Richard E., Krueger, David
Format:	Preprint
Published:	2025
Subjects:	Machine Learning Artificial Intelligence Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Implicit meta-learning may lead language models to trust more reliable sources
by: Krasheninnikov, Dmitrii, et al.
Published: (2023)

Stress-Testing Capability Elicitation With Password-Locked Models
by: Greenblatt, Ryan, et al.
Published: (2024)

Linguistic properties and model scale in brain encoding: from small to compressed language models
by: Oota, Subba Reddy, et al.
Published: (2026)

Vocabulary shapes cross-lingual variation of word-order learnability in language models
by: Martins, Jonas Mayer, et al.
Published: (2026)

How Do LLMs Persuade? Linear Probes Can Uncover Persuasion Dynamics in Multi-Turn Conversations
by: Jaipersaud, Brandon, et al.
Published: (2025)

The language of time: a language model perspective on time-series foundation models
by: Xie, Yi, et al.
Published: (2025)

Alignment faking in large language models
by: Greenblatt, Ryan, et al.
Published: (2024)

Representation in large language models
by: Yetman, Cameron
Published: (2025)

Auditing language models for hidden objectives
by: Marks, Samuel, et al.
Published: (2025)

Lightweight reranking for language model generations
by: Jain, Siddhartha, et al.
Published: (2023)

Long-form factuality in large language models
by: Wei, Jerry, et al.
Published: (2024)

Can large language models explore in-context?
by: Krishnamurthy, Akshay, et al.
Published: (2024)

Extracting books from production language models
by: Ahmed, Ahmed, et al.
Published: (2026)

Direction-Flipped Influence Audits Reveal Hidden Structure in Moral Choices of LLMs
by: Blandfort, Phil, et al.
Published: (2026)

Scaling laws for language encoding models in fMRI
by: Antonello, Richard, et al.
Published: (2023)

AI-AI Bias: large language models favor communications generated by large language models
by: Laurito, Walter, et al.
Published: (2024)

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

Badllama 3: removing safety finetuning from Llama 3 in minutes
by: Volkov, Dmitrii
Published: (2024)

Quantifying Feature Space Universality Across Large Language Models via Sparse Autoencoders
by: Lan, Michael, et al.
Published: (2024)

Post-training makes large language models less human-like
by: Binz, Marcel, et al.
Published: (2026)

Open-sci-ref-0.01: open and reproducible reference baselines for language model and dataset comparison
by: Nezhurina, Marianna, et al.
Published: (2025)

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

Clinical ModernBERT: An efficient and long context encoder for biomedical text
by: Lee, Simon A., et al.
Published: (2025)

Pretraining with hierarchical memories: separating long-tail and common knowledge
by: Pouransari, Hadi, et al.
Published: (2025)

Block removal for large language models through constrained binary optimization
by: Jansen, David, et al.
Published: (2026)

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

Efficient semantic uncertainty quantification in language models via diversity-steered sampling
by: Park, Ji Won, et al.
Published: (2025)

Safety and accuracy follow different scaling laws in clinical large language models
by: Wind, Sebastian, et al.
Published: (2026)

A dataset and benchmark for hospital course summarization with adapted large language models
by: Aali, Asad, et al.
Published: (2024)

CogBench: a large language model walks into a psychology lab
by: Coda-Forno, Julian, et al.
Published: (2024)

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

Multi-step retrieval and reasoning improves radiology question answering with large language models
by: Wind, Sebastian, et al.
Published: (2025)

MAP's not dead yet: Uncovering true language model modes by conditioning away degeneracy
by: Yoshida, Davis, et al.
Published: (2023)

SteuerLLM: Local specialized large language model for German tax law analysis
by: Wind, Sebastian, et al.
Published: (2026)

Multimodal large language model for wheat breeding: a new exploration of smart breeding
by: Yang, Guofeng, et al.
Published: (2024)

MedicalBERT: enhancing biomedical natural language processing using pretrained BERT-based model
by: Reddy, K. Sahit, et al.
Published: (2025)

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

Data Science Kitchen at GermEval 2021: A Fine Selection of Hand-Picked Features, Delivered Fresh from the Oven
by: Hildebrandt, Niclas, et al.
Published: (2021)

Can large language models replace humans in the systematic review process? Evaluating GPT-4's efficacy in screening and extracting data from peer-reviewed and grey literature in multiple languages
by: Khraisha, Qusai, et al.
Published: (2023)

TransformerFAM: Feedback attention is working memory
by: Hwang, Dongseong, et al.
Published: (2024)