Image de couverture de livre

Layerwise Recall and the Geometry of Interwoven Knowledge in LLMs

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Lei, Ge, Cooper, Samuel J.
Format:	Preprint
Publié:	2025
Sujets:	Computation and Language Artificial Intelligence Machine Learning
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Towards a Holistic Evaluation of LLMs on Factual Knowledge Recall
par: Yuan, Jiaqing, et autres
Publié: (2024)

Layerwise Change of Knowledge in Neural Networks
par: Cheng, Xu, et autres
Publié: (2024)

Predictable Confabulations: Factual Recall by LLMs Scales with Model Size and Topic Frequency
par: Smith, Matthew L., et autres
Publié: (2026)

Assessing Episodic Memory in LLMs with Sequence Order Recall Tasks
par: Pink, Mathis, et autres
Publié: (2024)

Outlier-weighed Layerwise Sampling for LLM Fine-tuning
par: Li, Pengxiang, et autres
Publié: (2024)

Locate-then-edit for Multi-hop Factual Recall under Knowledge Editing
par: Zhang, Zhuoran, et autres
Publié: (2024)

From Compression to Expression: A Layerwise Analysis of In-Context Learning
par: Jiang, Jiachen, et autres
Publié: (2025)

Censored LLMs as a Natural Testbed for Secret Knowledge Elicitation
par: Casademunt, Helena, et autres
Publié: (2026)

Certifying Knowledge Comprehension in LLMs
par: Chaudhary, Isha, et autres
Publié: (2024)

Interpreting Key Mechanisms of Factual Recall in Transformer-Based Language Models
par: Lv, Ang, et autres
Publié: (2024)

LongRecall: A Structured Approach for Robust Recall Evaluation in Long-Form Text
par: Ardestani, MohamamdJavad, et autres
Publié: (2025)

KSOD: Knowledge Supplement for LLMs On Demand
par: Li, Haoran, et autres
Publié: (2025)

StateX: Enhancing RNN Recall via Post-training State Expansion
par: Shen, Xingyu, et autres
Publié: (2025)

PropMEND: Hypernetworks for Knowledge Propagation in LLMs
par: Liu, Zeyu Leo, et autres
Publié: (2025)

LISA: Layerwise Importance Sampling for Memory-Efficient Large Language Model Fine-Tuning
par: Pan, Rui, et autres
Publié: (2024)

StepFun-Formalizer: Unlocking the Autoformalization Potential of LLMs through Knowledge-Reasoning Fusion
par: Wu, Yutong, et autres
Publié: (2025)

From Associations to Activations: Comparing Behavioral and Hidden-State Semantic Geometry in LLMs
par: Schiekiera, Louis, et autres
Publié: (2026)

Memorization vs. Reasoning: Updating LLMs with New Knowledge
par: Li, Aochong Oliver, et autres
Publié: (2025)

Fine-Tuning or Retrieval? Comparing Knowledge Injection in LLMs
par: Ovadia, Oded, et autres
Publié: (2023)

Adaptive Large Language Models By Layerwise Attention Shortcuts
par: Verma, Prateek, et autres
Publié: (2024)

RAG in the Wild: On the (In)effectiveness of LLMs with Mixture-of-Knowledge Retrieval Augmentation
par: Xu, Ran, et autres
Publié: (2025)

Recall-Extend Dynamics: Enhancing Small Language Models through Controlled Exploration and Refined Offline Integration
par: Guan, Zhong, et autres
Publié: (2025)

Your Absorbing Discrete Diffusion Secretly Models the Bayesian Posterior
par: Doyle, Cooper
Publié: (2025)

Everything Everywhere All at Once: LLMs can In-Context Learn Multiple Tasks in Superposition
par: Xiong, Zheyang, et autres
Publié: (2024)

Synergizing LLMs and Knowledge Graphs: A Novel Approach to Software Repository-Related Question Answering
par: Abedu, Samuel, et autres
Publié: (2024)

Reasoning's Razor: Reasoning Improves Accuracy but Can Hurt Recall at Critical Operating Points in Safety and Hallucination Detection
par: Chegini, Atoosa, et autres
Publié: (2025)

Data Doping or True Intelligence? Evaluating the Transferability of Injected Knowledge in LLMs
par: Jan, Essa, et autres
Publié: (2025)

Examining LLMs' Uncertainty Expression Towards Questions Outside Parametric Knowledge
par: Liu, Genglin, et autres
Publié: (2023)

Knowledge Localization in Mixture-of-Experts LLMs Using Cross-Lingual Inconsistency
par: Bandarkar, Lucas, et autres
Publié: (2026)

The Reversal Curse: LLMs trained on "A is B" fail to learn "B is A"
par: Berglund, Lukas, et autres
Publié: (2023)

Automated test generation to evaluate tool-augmented LLMs as conversational AI agents
par: Arcadinho, Samuel, et autres
Publié: (2024)

Frontier LLMs Still Struggle with Simple Reasoning Tasks
par: Malek, Alan, et autres
Publié: (2025)

KARL: Mitigating Hallucinations in LLMs via Knowledge-Boundary-Aware Reinforcement Learning
par: Gao, Cheng, et autres
Publié: (2026)

ContextGPT: Infusing LLMs Knowledge into Neuro-Symbolic Activity Recognition Models
par: Arrotta, Luca, et autres
Publié: (2024)

Structured Extraction of Real World Medical Knowledge using LLMs for Summarization and Search
par: Kim, Edward, et autres
Publié: (2024)

RECALL-MM: A Multimodal Dataset of Consumer Product Recalls for Risk Analysis using Computational Methods and Large Language Models
par: Bolanos, Diana, et autres
Publié: (2025)

Enhancing Delta Compression in LLMs via SVD-based Quantization Error Minimization
par: Xiong, Boya, et autres
Publié: (2025)

Project Alexandria: Towards Freeing Scientific Knowledge from Copyright Burdens via LLMs
par: Schuhmann, Christoph, et autres
Publié: (2025)

KDRL: Post-Training Reasoning LLMs via Unified Knowledge Distillation and Reinforcement Learning
par: Xu, Hongling, et autres
Publié: (2025)

Principled Fine-tuning of LLMs from User-Edits: A Medley of Preference, Supervision, and Reward
par: Misra, Dipendra, et autres
Publié: (2026)