paper-with-me

Papers

Sparse Autoencoders Can Interpret Randomly Initialized Transformers

2025-01-29 · Thomas Heap, Tim Lawson, Lucy Farnik, Laurence Aitchison

Sparse autoencoders (SAEs) are an increasingly popular technique for interpreting the internal representations of transformers. In this paper, we apply SAEs to 'interpret' random transformers, i.e., transformers where the parameters are sampled IID from a Gaussian rather than trained on text data. We find that random and trained transformers produce similarly interpretable SAE latents, and we confirm this finding quantitatively using an open-source auto-interpretability pipeline. Further, we find that SAE quality metrics are broadly similar for random and trained transformers. We find that these results hold across model sizes and layers. We discuss a number of number interesting questions that this work raises for the use of SAEs and auto-interpretability in the context of mechanistic interpretability.

📄 PDF Abstract BibTeX arXiv:2501.17727

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Algorithmic Capabilities of Random Transformers

2024-10-06 · Ziqian Zhong, Jacob Andreas

Trained transformer models have been found to implement interpretable procedures for tasks like arithmetic and associative recall, but little is understood about how the circuits that implement these procedures originate…

Text Generation

TIDE : Temporal-Aware Sparse Autoencoders for Interpretable Diffusion Transformers in Image Generation

2025-03-10 · Victor Shea-Jay Huang, Le Zhuo, Yi Xin, Zhaokai Wang 외

Diffusion Transformers (DiTs) are a powerful yet underexplored class of generative models compared to U-Net-based diffusion models. To bridge this gap, we introduce TIDE (Temporal-aware Sparse Autoencoders for Interpreta…

DenoisingImage GenerationStyle Transfer

Fixed Universal Transformers

2026-05-29 · Jingwen Liu, Alexandr Andoni, Daniel Hsu arxiv

We introduce \emph{universal transformers}: fixed transformers that can simulate any transformer in a given class via a suitable input embedding. Analogous to a universal Turing machine, the input embedding encodes a des…

Attribution-Guided Distillation of Matryoshka Sparse Autoencoders

2025-12-31 · Cristina P. Martin-Linares, Jonathan P. Ling arxiv

Sparse autoencoders (SAEs) aim to disentangle model activations into monosemantic, human-interpretable features. In practice, learned features are often redundant and vary across training runs and sparsity levels, which …

Interpreting Attention Layer Outputs with Sparse Autoencoders

2024-06-25 · Connor Kissane, Robert Krzyzanowski, Joseph Isaac Bloom, Arthur Conmy 외

Decomposing model activations into interpretable components is a key open problem in mechanistic interpretability. Sparse autoencoders (SAEs) are a popular method for decomposing the internal activations of trained trans…