Cover Image

Weight Patching: Toward Source-Level Mechanistic Localization in LLMs

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Sun, Chenghao, Zhang, Chengsheng, Qin, Guanzheng, Dai, Rui, Tian, Xinmei
Format:	Preprint
Published:	2026
Subjects:	Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Interpreting and Enhancing Emotional Circuits in Large Vision-Language Models via Cross-Modal Information Flow
by: Zhang, Chengsheng, et al.
Published: (2026)

Prefill-Time Intervention for Mitigating Hallucination in Large Vision-Language Models
by: Zhang, Chengsheng, et al.
Published: (2026)

MOSS: Self-Evolution through Source-Level Rewriting in Autonomous Agent Systems
by: Cai, Qianshu, et al.
Published: (2026)

Towards Theoretical Understandings of Self-Consuming Generative Models
by: Fu, Shi, et al.
Published: (2024)

MSCR: Exploring the Vulnerability of LLMs' Mathematical Reasoning Abilities Using Multi-Source Candidate Replacement
by: Sun, Zhishen, et al.
Published: (2025)

Localized Definitions and Distributed Reasoning: A Proof-of-Concept Mechanistic Interpretability Study via Activation Patching
by: Bahador, Nooshin
Published: (2025)

Drawback of Enforcing Equivariance and its Compensation via the Lens of Expressive Power
by: Chen, Yuzhu, et al.
Published: (2025)

LLMIA: An Out-of-the-Box Index Advisor via In-Context Learning with LLMs
by: Zhao, Xinxin, et al.
Published: (2025)

PKU-SafeRLHF: Towards Multi-Level Safety Alignment for LLMs with Human Preference
by: Ji, Jiaming, et al.
Published: (2024)

ATIR: Towards Audio-Text Interleaved Contextual Retrieval
by: Zhao, Tong, et al.
Published: (2026)

PatchFinder: A Two-Phase Approach to Security Patch Tracing for Disclosed Vulnerabilities in Open-Source Software
by: Li, Kaixuan, et al.
Published: (2024)

Intrinsic Self-Correction in LLMs: Towards Explainable Prompting via Mechanistic Interpretability
by: Lee, Yu-Ting, et al.
Published: (2025)

UAVs Meet LLMs: Overviews and Perspectives Toward Agentic Low-Altitude Mobility
by: Tian, Yonglin, et al.
Published: (2025)

ME-Mamba: Multi-Expert Mamba with Efficient Knowledge Capture and Fusion for Multimodal Survival Analysis
by: Zhang, Chengsheng, et al.
Published: (2025)

How Emotion Shapes the Behavior of LLMs and Agents: A Mechanistic Study
by: Sun, Moran, et al.
Published: (2026)

Beyond Surface-Level Detection: Towards Cognitive-Driven Defense Against Jailbreak Attacks via Meta-Operations Reasoning
by: Pu, Rui, et al.
Published: (2025)

Exploring Large Language Models for Knowledge Graph Completion
by: Yao, Liang, et al.
Published: (2023)

A Theoretical Perspective: How to Prevent Model Collapse in Self-consuming Training Loops
by: Fu, Shi, et al.
Published: (2025)

LocalBench: Benchmarking LLMs on County-Level Local Knowledge and Reasoning
by: Gao, Zihan, et al.
Published: (2025)

RankLLM: Weighted Ranking of LLMs by Quantifying Question Difficulty
by: Zhang, Ziqian, et al.
Published: (2026)

Dissecting Bias in LLMs: A Mechanistic Interpretability Perspective
by: Chandna, Bhavik, et al.
Published: (2025)

Expand Heterogeneous Learning Systems with Selective Multi-Source Knowledge Fusion
by: Dai, Gaole, et al.
Published: (2024)

APTBench: Benchmarking Agentic Potential of Base LLMs During Pre-Training
by: Qin, Jiarui, et al.
Published: (2025)

HyperDAS: Towards Automating Mechanistic Interpretability with Hypernetworks
by: Sun, Jiuding, et al.
Published: (2025)

Towards Autonomous Mechanistic Reasoning in Virtual Cells
by: Jang, Yunhui, et al.
Published: (2026)

How Post-Training Reshapes LLMs: A Mechanistic View on Knowledge, Truthfulness, Refusal, and Confidence
by: Du, Hongzhe, et al.
Published: (2025)

Hierarchical Repository-Level Code Summarization for Business Applications Using Local LLMs
by: Dhulshette, Nilesh, et al.
Published: (2025)

SUDO: a framework for evaluating clinical artificial intelligence systems without ground-truth annotations
by: Kiyasseh, Dani, et al.
Published: (2024)

D$^2$Quant: Accurate Low-bit Post-Training Weight Quantization for LLMs
by: Yan, Xianglong, et al.
Published: (2026)

Analysing Moral Bias in Finetuned LLMs through Mechanistic Interpretability
by: Raimondi, Bianca, et al.
Published: (2025)

Safety at One Shot: Patching Fine-Tuned LLMs with A Single Instance
by: Zhang, Jiawen, et al.
Published: (2026)

A Picture is Worth A Thousand Numbers: Enabling LLMs Reason about Time Series via Visualization
by: Liu, Haoxin, et al.
Published: (2024)

Patched RTC: evaluating LLMs for diverse software development tasks
by: Sharma, Asankhaya
Published: (2024)

ATime-Consistent Benchmark for Repository-Level Software Engineering Evaluation
by: Xianpeng, et al.
Published: (2026)

Towards Foundation Models for Zero-Shot Time Series Anomaly Detection: Leveraging Synthetic Data and Relative Context Discrepancy
by: Lan, Tian, et al.
Published: (2025)

Learning From Developers: Towards Reliable Patch Validation at Scale for Linux
by: Lin, Chih-En, et al.
Published: (2026)

Mechanistic Knobs in LLMs: Retrieving and Steering High-Order Semantic Features via Sparse Autoencoders
by: Zhang, Ruikang, et al.
Published: (2026)

The Struggle Between Continuation and Refusal: A Mechanistic Analysis of the Continuation-Triggered Jailbreak in LLMs
by: Deng, Yonghong, et al.
Published: (2026)

Rethinking Code Similarity for Automated Algorithm Design with LLMs
by: Zhang, Rui, et al.
Published: (2026)

Calibration and Transformation-Free Weight-Only LLMs Quantization via Dynamic Grouping
by: Zheng, Xinzhe, et al.
Published: (2025)