Cover Image

LLM-Microscope: Uncovering the Hidden Role of Punctuation in Context Memory of Transformers

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Razzhigaev, Anton, Mikhalchuk, Matvey, Rahmatullaev, Temurbek, Goncharova, Elizaveta, Druzhinina, Polina, Oseledets, Ivan, Kuznetsov, Andrey
Format:	Preprint
Published:	2025
Subjects:	Computation and Language Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Universal Adversarial Attack on Aligned Multimodal LLMs
by: Rahmatullaev, Temurbek, et al.
Published: (2025)

The Shape of Learning: Anisotropy and Intrinsic Dimensions in Transformer-Based Models
by: Razzhigaev, Anton, et al.
Published: (2023)

Your Transformer is Secretly Linear
by: Razzhigaev, Anton, et al.
Published: (2024)

SONAR-LLM: Autoregressive Transformer that Thinks in Sentence Embeddings and Speaks in Tokens
by: Dragunov, Nikita, et al.
Published: (2025)

Addressing Hallucinations in Language Models with Knowledge Graph Embeddings as an Additional Modality
by: Chekalina, Viktoriia, et al.
Published: (2024)

MindShift: Analyzing Language Models' Reactions to Psychological Prompts
by: Vasiliuk, Anton, et al.
Published: (2025)

OmniFusion Technical Report
by: Goncharova, Elizaveta, et al.
Published: (2024)

NoReGeo: Non-Reasoning Geometry Benchmark
by: Abdullaeva, Irina, et al.
Published: (2026)

Back to Basics: Revisiting Exploration in Reinforcement Learning for LLM Reasoning via Generative Probabilities
by: Li, Pengyi, et al.
Published: (2026)

I Have Covered All the Bases Here: Interpreting Reasoning Features in Large Language Models via Sparse Autoencoders
by: Galichin, Andrey, et al.
Published: (2025)

Inverting Black-Box Face Recognition Systems via Zero-Order Optimization in Eigenface Space
by: Razzhigaev, Anton, et al.
Published: (2025)

Real-World Transferable Adversarial Attack on Face-Recognition Systems
by: Kaznacheev, Andrey, et al.
Published: (2025)

Sentence-Anchored Gist Compression for Long-Context LLMs
by: Tarasov, Dmitrii, et al.
Published: (2025)

Logit-KL Flow Matching: Non-Autoregressive Text Generation via Sampling-Hybrid Inference
by: Sevriugov, Egor, et al.
Published: (2024)

Simple Vision-Language Math Reasoning via Rendered Text
by: Skripkin, Matvey, et al.
Published: (2025)

Heterogeneous LLM Methods for Ontology Learning (Few-Shot Prompting, Ensemble Typing, and Attention-Based Taxonomies)
by: Beliaeva, Aleksandra, et al.
Published: (2025)

Feature-Level Insights into Artificial Text Detection with Sparse Autoencoders
by: Kuznetsov, Kristian, et al.
Published: (2025)

MOVE: A Mixture-of-Vision-Encoders Approach for Domain-Focused Vision-Language Processing
by: Skripkin, Matvey, et al.
Published: (2025)

Confidence Is All You Need: Few-Shot RL Fine-Tuning of Language Models
by: Li, Pengyi, et al.
Published: (2025)

Exploring the Hidden Capacity of LLMs for One-Step Text Generation
by: Mezentsev, Gleb, et al.
Published: (2025)

On the Spatial Structure of Mixture-of-Experts in Transformers
by: Bershatsky, Daniel, et al.
Published: (2025)

Listener-Rewarded Thinking in VLMs for Image Preferences
by: Gambashidze, Alexander, et al.
Published: (2025)

When Less is Enough: Adaptive Token Reduction for Efficient Image Representation
by: Allakhverdov, Eduard, et al.
Published: (2025)

Multi-Agent GraphRAG: A Text-to-Cypher Framework for Labeled Property Graphs
by: Gusarov, Anton, et al.
Published: (2025)

Polyhedral products, graph products and p-central series
by: Panov, Taras, et al.
Published: (2024)

Diagonal Batching Unlocks Parallelism in Recurrent Memory Transformers for Long Contexts
by: Sivtsov, Danil, et al.
Published: (2025)

MERA: A Comprehensive LLM Evaluation in Russian
by: Fenogenova, Alena, et al.
Published: (2024)

Longer Context, Deeper Thinking: Uncovering the Role of Long-Context Ability in Reasoning
by: Yang, Wang, et al.
Published: (2025)

Test-Time Reasoning Through Visual Human Preferences with VLMs and Soft Rewards
by: Gambashidze, Alexander, et al.
Published: (2025)

Image Reconstruction as a Tool for Feature Analysis
by: Allakhverdov, Eduard, et al.
Published: (2025)

FLAME: Flexible LLM-Assisted Moderation Engine
by: Bakulin, Ivan, et al.
Published: (2025)

CoMa: Contextual Massing Generation with Vision-Language Models
by: Maslov, Evgenii, et al.
Published: (2026)

MLLM-Microscope: Unlocking Hidden Structure Within Multimodal Large Language Models
by: Mussabayev, Ravil, et al.
Published: (2026)

Layer by Layer: Uncovering Hidden Representations in Language Models
by: Skean, Oscar, et al.
Published: (2025)

Scaling BERT Models for Turkish Automatic Punctuation and Capitalization Correction
by: Saoud, Abdulkader, et al.
Published: (2024)

Uncovering Hidden Intentions: Exploring Prompt Recovery for Deeper Insights into Generated Texts
by: Give, Louis, et al.
Published: (2024)

When Punctuation Matters: A Large-Scale Comparison of Prompt Robustness Methods for LLMs
by: Seleznyov, Mikhail, et al.
Published: (2025)

Resolving Transcription Ambiguity in Spanish: A Hybrid Acoustic-Lexical System for Punctuation Restoration
by: Zhu, Xiliang, et al.
Published: (2024)

RCR-Router: Efficient Role-Aware Context Routing for Multi-Agent LLM Systems with Structured Memory
by: Liu, Jun, et al.
Published: (2025)

Punctuation Prediction for Polish Texts using Transformers
by: Pokrywka, Jakub
Published: (2024)