paper-with-me

홈 › Papers

LLM Interpretability with Identifiable Temporal-Instantaneous Representation

2025-09-27 · Xiangchen Song, Jiaqi Sun, Zijian Li, Yujia Zheng, Kun Zhang arxiv

Despite Large Language Models' remarkable capabilities, understanding their internal representations remains challenging. Mechanistic interpretability tools such as sparse autoencoders (SAEs) were developed to extract interpretable features from LLMs but lack temporal dependency modeling, instantaneous relation representation, and more importantly theoretical guarantees, undermining both the theoretical foundations and the practical confidence necessary for subsequent analyses. While causal representation learning (CRL) offers theoretically grounded approaches for uncovering latent concepts, existing methods cannot scale to LLMs' rich conceptual space due to inefficient computation. To bridge the gap, we introduce an identifiable temporal causal representation learning framework specifically designed for LLMs' high-dimensional concept space, capturing both time-delayed and instantaneous causal relations. Our approach provides theoretical guarantees and demonstrates efficacy on synthetic datasets scaled to match real-world complexity. By extending SAE techniques with our temporal causal framework, we successfully discover meaningful concept relationships in LLM activations. Our findings show that modeling both temporal and instantaneous conceptual relationships advances the interpretability of LLMs.

📄 PDF Abstract BibTeX arXiv:2509.23323

Code (0)

등록된 구현이 없습니다.

Tasks

Representation Learning

Similar Papers 제목 키워드 기반

MOSAIC: Module Discovery via Sparse Additive Identifiable Causal Learning for Scientific Time Series

2026-05-06 · Shicheng Fan, Nour Elhendawy, Jianle Sun, Ke Fang 외 arxiv

Causal representation learning (CRL) seeks to recover latent variables with identifiability guarantees, typically up to permutation and component-wise reparameterization under appropriate assumptions. However, identifiab…

Representation Learning

Identifiable Markov Switching Models with Instantaneous Effects and Exponential Families

2026-06-01 · Roel Hulsman, Carles Balsells-Rodas, Sara Magliacane arxiv

Temporal systems often exhibit non-stationary behaviour, such as seasonal climate variation or glucose fluctuations in patients with type-1 diabetes. One way to model non-stationarity is through discrete latent regimes, …

TRACE: Trajectory Recovery for Continuous Mechanism Evolution in Causal Representation Learning

2026-01-29 · Shicheng Fan, Kun Zhang, Lu Cheng arxiv

Temporal causal representation learning methods assume that causal mechanisms switch instantaneously between discrete domains, yet real-world systems often exhibit continuous mechanism transitions. For example, a vehicle…

Representation Learning

Causal Representation Learning for Instantaneous and Temporal Effects in Interactive Systems

2022-06-13 · Phillip Lippe, Sara Magliacane, Sindy Löwe, Yuki M. Asano 외

Causal representation learning is the task of identifying the underlying causal variables and their relations from high-dimensional observations, such as images. Recent work has shown that one can reconstruct the causal …

Causal DiscoveryRepresentation LearningTemporal Sequences

Interpreting Video Representations with Spatio-Temporal Sparse Autoencoders

2026-04-05 · Atahan Dokme, Sriram Vishwanath arxiv

We present the first systematic study of Sparse Autoencoders (SAEs) on video representations. Standard SAEs decompose video into interpretable, monosemantic features but destroy temporal coherence: hard TopK selection pr…

Action ClassificationVideo Retrieval