Copertina

All Circuits Lead to Rome: Rethinking Functional Anisotropy in Circuit and Sheaf Discovery for LLMs

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Chen, Xi, Jin, Mingyu, Niu, Jingcheng, Yin, Yutong, Zhao, Jinman, Guo, Bangwei, Metaxas, Dimitris N., Wang, Zhaoran, Yue, Yutao, Penn, Gerald
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Computation and Language
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Sheaf Discovery with Joint Computation Graph Pruning and Flexible Granularity
di: Yu, Lei, et al.
Pubblicazione: (2024)

What Happens Inside Agent Memory? Circuit Analysis from Emergence to Diagnosis
di: Mao, Xutao, et al.
Pubblicazione: (2026)

Farther the Shift, Sparser the Representation: Analyzing OOD Mechanisms in LLMs
di: Jin, Mingyu, et al.
Pubblicazione: (2026)

What does the Knowledge Neuron Thesis Have to do with Knowledge?
di: Niu, Jingcheng, et al.
Pubblicazione: (2024)

Learning Volumetric Neural Deformable Models to Recover 3D Regional Heart Wall Motion from Multi-Planar Tagged MRI
di: Ye, Meng, et al.
Pubblicazione: (2024)

Rethinking Catheters to Reduce Costs in Urodynamic Evaluation: Do All Roads Lead to Rome?
di: João Antonio Pereira‐Correia, et al.
Pubblicazione: (2026)

Are Transformers Able to Reason by Connecting Separated Knowledge in Training Data?
di: Yin, Yutong, et al.
Pubblicazione: (2025)

VerSe: Integrating Multiple Queries as Prompts for Versatile Cardiac MRI Segmentation
di: Guo, Bangwei, et al.
Pubblicazione: (2024)

LUCID-SAE: Learning Unified Vision-Language Sparse Codes for Interpretable Concept Discovery
di: Gu, Difei, et al.
Pubblicazione: (2026)

All Roads Lead to Rome: Comparing Nanoparticle‐ and Small Molecule‐Driven Cell Autophagy
di: Xiaofei Zhou, et al.
Pubblicazione: (2024)

All Roads Lead to Rome: Let's Choose the Safest One!
di: Francesco Catania, et al.
Pubblicazione: (2025)

Do LLMs "Feel"? Emotion Circuits Discovery and Control
di: Wang, Chenxi, et al.
Pubblicazione: (2025)

Understanding Verbatim Memorization in LLMs Through Circuit Discovery
di: Lasy, Ilya, et al.
Pubblicazione: (2025)

M^3-Bench: Multi-Modal, Multi-Hop, Multi-Threaded Tool-Using MLLM Agent Benchmark
di: Zhou, Yang, et al.
Pubblicazione: (2025)

GeoSheaf: A Sheaf-Theoretic Framework for Seismic-Well Integration
di: Gerald Enrique Nelson Mc Kenzie
Pubblicazione: (2025)

K-Prism: A Knowledge-Guided and Prompt Integrated Universal Medical Image Segmentation Model
di: Guo, Bangwei, et al.
Pubblicazione: (2025)

All Roads Lead to Rome: Pathways to Engineering Disease Resistance in Plants
di: Aziz Ul Ikram, et al.
Pubblicazione: (2024)

$λ$-GRPO: Unifying the GRPO Frameworks with Learnable Token Preferences
di: Wang, Yining, et al.
Pubblicazione: (2025)

MPDiT: Multi-Patch Global-to-Local Transformer Architecture For Efficient Flow Matching and Diffusion Model
di: Dao, Quan, et al.
Pubblicazione: (2026)

All Roads Lead to Rome: Incentivizing Divergent Thinking in Vision-Language Models
di: Tian, Xinyu, et al.
Pubblicazione: (2026)

Automatic Discovery of Visual Circuits
di: Rajaram, Achyuta, et al.
Pubblicazione: (2024)

Rethinking Probabilistic Circuit Parameter Learning
di: Liu, Anji, et al.
Pubblicazione: (2025)

IBCircuit: Towards Holistic Circuit Discovery with Information Bottleneck
di: Bian, Tian, et al.
Pubblicazione: (2026)

Can LLMs Predict Their Own Failures? Self-Awareness via Internal Circuits
di: Ghasemabadi, Amirhosein, et al.
Pubblicazione: (2025)

All Roads Lead to Rome: Unveiling the Trajectory of Recommender Systems Across the LLM Era
di: Chen, Bo, et al.
Pubblicazione: (2024)

Token-Controlled Re-ranking for Sequential Recommendation via LLMs
di: Dai, Wenxi, et al.
Pubblicazione: (2025)

Beyond Explicit Edges: Robust Reasoning over Noisy and Sparse Knowledge Graphs
di: Gao, Hang, et al.
Pubblicazione: (2026)

Position-aware Automatic Circuit Discovery
di: Haklay, Tal, et al.
Pubblicazione: (2025)

Spectra: Rethinking Optimizers for LLMs Under Spectral Anisotropy
di: Huang, Zhendong, et al.
Pubblicazione: (2026)

All Roads Lead to Rome: Graph-Based Confidence Estimation for Large Language Model Reasoning
di: Zhang, Caiqi, et al.
Pubblicazione: (2025)

RealRoute: Dynamic Query Routing System via Retrieve-then-Verify Paradigm
di: Liu, Jiahe, et al.
Pubblicazione: (2026)

Granular Concept Circuits: Toward a Fine-Grained Circuit Discovery for Concept Representations
di: Kwon, Dahee, et al.
Pubblicazione: (2025)

Circuit Tracing in Vision-Language Models: Understanding the Internal Mechanisms of Multimodal Thinking
di: Yang, Jingcheng, et al.
Pubblicazione: (2026)

Rethinking Circuit Completeness in Language Models: AND, OR, and ADDER Gates
di: Chen, Hang, et al.
Pubblicazione: (2025)

Target the Circuit, Not the Brain: Rethinking Essential Tremor Therapy
di: Ananda Carolina Moraes de Falcone, et al.
Pubblicazione: (2025)

The Computational Complexity of Circuit Discovery for Inner Interpretability
di: Adolfi, Federico, et al.
Pubblicazione: (2024)

Multi-Granular Node Pruning for Circuit Discovery
di: Haider, Muhammad Umair, et al.
Pubblicazione: (2025)

All Roads Lead to Rome? Exploring Representational Similarities Between Latent Spaces of Generative Image Models
di: Badrinath, Charumathi, et al.
Pubblicazione: (2024)

If All Roads Lead to Rome, Why Are We in Athens?: School Library Media Programs and National Standards.
di: Barron, Daniel D.
Pubblicazione: (1997)

Test-Time Spectrum-Aware Latent Steering for Zero-Shot Generalization in Vision-Language Models
di: Dafnis, Konstantinos M., et al.
Pubblicazione: (2025)