paper-with-me

Papers

InverseScope: Scalable Activation Inversion for Interpreting Large Language Models

2025-06-09 · Yifan Luo, Zhennan Zhou, Bin Dong arxiv

Understanding the internal representations of large language models (LLMs) is a central challenge in interpretability research. Existing feature interpretability methods often rely on strong structural assumptions--such as linearity or sparsity--that may not hold in practice. In this work, we introduce InverseScope, an assumption-light and scalable framework for interpreting neural activations via input inversion. Given a target activation, InverseScope characterizes its encoded information by generating natural-language inputs that produce nearby activations, grounding abstract internal states in concrete language. To overcome the prohibitive cost of sampling in high-dimensional activation spaces, we propose a novel control-layer conditioning architecture that substantially improves sample efficiency compared to prior token-prepending approaches. We demonstrate that InverseScope reveals rich geometric structure in LLM representation spaces, including sentence-level linear analogies. The framework scales to state-of-the-art open-source models of up to 14B parameters and generalizes to out-of-distribution inputs, enabling systematic analysis of activation neighborhoods.

📄 PDF Abstract BibTeX arXiv:2506.07406

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Building Better Activation Oracles

2026-05-23 · Jan Bauer, Celeste De Schamphelaere, Adam Karvonen, Niclas Luick 외 arxiv

Activation Oracles (AOs) are promising methods for interpreting residual stream activations. However, current AOs face important issues, such as hallucinations and vagueness. Additionally, text-inversion confounds make t…

ARES: Scalable and Practical Gradient Inversion Attack in Federated Learning through Activation Recovery

2026-03-18 · Zirui Gong, Leo Yu Zhang, Yanjun Zhang, Viet Vo 외 arxiv

Federated Learning (FL) enables collaborative model training by sharing model updates instead of raw data, aiming to protect user privacy. However, recent studies reveal that these shared updates can inadvertently leak s…

Federated Learning

Predictive Concept Decoders: Training Scalable End-to-End Interpretability Assistants

2025-12-17 · Vincent Huang, Dami Choi, Daniel D. Johnson, Sarah Schwettmann 외 arxiv

Interpreting the internal activations of neural networks can produce more faithful explanations of their behavior, but is difficult due to the complex structure of activation space. Existing approaches to scalable interp…

N2G: A Scalable Approach for Quantifying Interpretable Neuron Representations in Large Language Models

2023-04-22 · Alex Foote, Neel Nanda, Esben Kran, Ionnis Konstas 외

Understanding the function of individual neurons within language models is essential for mechanistic interpretability research. We propose $\textbf{Neuron to Graph (N2G)}$, a tool which takes a neuron and its dataset exa…

Inference-Time Decomposition of Activations (ITDA): A Scalable Approach to Interpreting Large Language Models

2025-05-23 · Patrick Leask, Neel Nanda, Noura Al Moubayed

Sparse autoencoders (SAEs) are a popular method for decomposing Large Langage Models (LLM) activations into interpretable latents. However, due to their substantial training cost, most academic research uses open-source …

GPULanguage ModelingLanguage Modelling