paper-with-me

Papers

How LLMs Learn: Tracing Internal Representations with Sparse Autoencoders

2025-03-09 · Tatsuro Inaba, Kentaro Inui, Yusuke Miyao, Yohei Oseki, Benjamin Heinzerling, Yu Takagi

Large Language Models (LLMs) demonstrate remarkable multilingual capabilities and broad knowledge. However, the internal mechanisms underlying the development of these capabilities remain poorly understood. To investigate this, we analyze how the information encoded in LLMs' internal representations evolves during the training process. Specifically, we train sparse autoencoders at multiple checkpoints of the model and systematically compare the interpretative results across these stages. Our findings suggest that LLMs initially acquire language-specific knowledge independently, followed by cross-linguistic correspondences. Moreover, we observe that after mastering token-level knowledge, the model transitions to learning higher-level, abstract concepts, indicating the development of more conceptual understanding.

📄 PDF Abstract BibTeX arXiv:2503.06394

Code (1)

llm-jp/llm-jp-sae 공식 구현 pytorch

Similar Papers 제목 키워드 기반

Tracing Moral Foundations in Large Language Models

2026-01-09 · Chenxiao Yu, Bowen Yi, Farzan Karimi-Malekabadi, Suhaib Abdurahman 외 arxiv

Large language models often produce human-like moral judgments, but it is unclear whether this reflects an internal conceptual structure or superficial ``moral mimicry.'' Using Moral Foundations Theory (MFT) as an analyt…

CulTrace: Tracing Internal Cultural Reasoning in Large Language Models

2025-08-12 · Haeun Yu, Seogyeong Jeong, Siddhesh Pawar, Jisu Shin 외 arxiv

The growing deployment of large language models (LLMs) across diverse cultural contexts necessitates a deeper understanding of models' hidden representations of different cultures. Prior work has evaluated cultural aware…

Finding Neurons in a Haystack: Case Studies with Sparse Probing

2023-05-02 · Wes Gurnee, Neel Nanda, Matthew Pauly, Katherine Harvey 외

Despite rapid adoption and deployment of large language models (LLMs), the internal computations of these models remain opaque and poorly understood. In this work, we seek to understand how high-level human-interpretable…

Farther the Shift, Sparser the Representation: Analyzing OOD Mechanisms in LLMs

2026-03-03 · Mingyu Jin, Yutong Yin, Jingcheng Niu, Qingcheng Zeng 외 arxiv

In this work, we investigate how Large Language Models (LLMs) adapt their internal representations when encountering inputs of increasing difficulty, quantified as the degree of out-of-distribution (OOD) shift. We reveal…

The Quantum Sieve Tracer: A Hybrid Framework for Layer-Wise Activation Tracing in Large Language Models

2026-02-06 · Jonathan Pan arxiv

Mechanistic interpretability aims to reverse-engineer the internal computations of Large Language Models (LLMs), yet separating sparse semantic signals from high-dimensional polysemantic noise remains a significant chall…