Compute Optimal Inference and Provable Amortisation Gap in Sparse Autoencoders

• URL: http://arxiv.org/abs/2411.13117v1
• Date: Wed, 20 Nov 2024 08:21:53 GMT
• Title: Compute Optimal Inference and Provable Amortisation Gap in Sparse Autoencoders
• Authors: Charles O'Neill, David Klindt,
• Abstract summary: We investigate sparse inference and learning in SAEs through the lens of sparse coding. We show that SAEs perform amortised sparse inference with a computationally restricted encoder. We empirically explore conditions where more sophisticated sparse inference methods outperform traditional SAE encoders.
• Score: 0.0
• License:
• Abstract: A recent line of work has shown promise in using sparse autoencoders (SAEs) to uncover interpretable features in neural network representations. However, the simple linear-nonlinear encoding mechanism in SAEs limits their ability to perform accurate sparse inference. In this paper, we investigate sparse inference and learning in SAEs through the lens of sparse coding. Specifically, we show that SAEs perform amortised sparse inference with a computationally restricted encoder and, using compressed sensing theory, we prove that this mapping is inherently insufficient for accurate sparse inference, even in solvable cases. Building on this theory, we empirically explore conditions where more sophisticated sparse inference methods outperform traditional SAE encoders. Our key contribution is the decoupling of the encoding and decoding processes, which allows for a comparison of various sparse encoding strategies. We evaluate these strategies on two dimensions: alignment with true underlying sparse features and correct inference of sparse codes, while also accounting for computational costs during training and inference. Our results reveal that substantial performance gains can be achieved with minimal increases in compute cost. We demonstrate that this generalises to SAEs applied to large language models (LLMs), where advanced encoders achieve similar interpretability. This work opens new avenues for understanding neural network representations and offers important implications for improving the tools we use to analyse the activations of large language models.

Related papers

• A Theoretical Perspective for Speculative Decoding Algorithm [60.79447486066416]
  One effective way to accelerate inference is emphSpeculative Decoding, which employs a small model to sample a sequence of draft tokens and a large model to validate. This paper tackles this gap by conceptualizing the decoding problem via markov chain abstraction and studying the key properties, emphoutput quality and inference acceleration, from a theoretical perspective.
  arXiv  Detail & Related papers  (2024-10-30T01:53:04Z)
• Interpretability as Compression: Reconsidering SAE Explanations of Neural Activations with MDL-SAEs [0.0]
  We present an information-theoretic framework for interpreting SAEs as lossy compression algorithms. We argue that using MDL rather than sparsity may avoid potential pitfalls with naively maximising sparsity.
  arXiv  Detail & Related papers  (2024-10-15T01:38:03Z)
• Disentangling Dense Embeddings with Sparse Autoencoders [0.0]
  Sparse autoencoders (SAEs) have shown promise in extracting interpretable features from complex neural networks. We present one of the first applications of SAEs to dense text embeddings from large language models. We show that the resulting sparse representations maintain semantic fidelity while offering interpretability.
  arXiv  Detail & Related papers  (2024-08-01T15:46:22Z)
• Symmetric Equilibrium Learning of VAEs [56.56929742714685]
  We view variational autoencoders (VAEs) as decoder-encoder pairs, which map distributions in the data space to distributions in the latent space and vice versa. We propose a Nash equilibrium learning approach, which is symmetric with respect to the encoder and decoder and allows learning VAEs in situations where both the data and the latent distributions are accessible only by sampling.
  arXiv  Detail & Related papers  (2023-07-19T10:27:34Z)
• In-context Autoencoder for Context Compression in a Large Language Model [70.7621953091318]
  We propose the In-context Autoencoder (ICAE) to compress a long context into short compact memory slots. ICAE is first pretrained using both autoencoding and language modeling objectives on massive text data.
  arXiv  Detail & Related papers  (2023-07-13T17:59:21Z)
• Improving Deep Representation Learning via Auxiliary Learnable Target Coding [69.79343510578877]
  This paper introduces a novel learnable target coding as an auxiliary regularization of deep representation learning. Specifically, a margin-based triplet loss and a correlation consistency loss on the proposed target codes are designed to encourage more discriminative representations.
  arXiv  Detail & Related papers  (2023-05-30T01:38:54Z)
• Fundamental Limits of Two-layer Autoencoders, and Achieving Them with Gradient Methods [91.54785981649228]
  This paper focuses on non-linear two-layer autoencoders trained in the challenging proportional regime. Our results characterize the minimizers of the population risk, and show that such minimizers are achieved by gradient methods. For the special case of a sign activation function, our analysis establishes the fundamental limits for the lossy compression of Gaussian sources via (shallow) autoencoders.
  arXiv  Detail & Related papers  (2022-12-27T12:37:34Z)
• Variational Sparse Coding with Learned Thresholding [6.737133300781134]
  We propose a new approach to variational sparse coding that allows us to learn sparse distributions by thresholding samples. We first evaluate and analyze our method by training a linear generator, showing that it has superior performance, statistical efficiency, and gradient estimation.
  arXiv  Detail & Related papers  (2022-05-07T14:49:50Z)
• The Interpretable Dictionary in Sparse Coding [4.205692673448206]
  In our work, we illustrate that an ANN, trained using sparse coding under specific sparsity constraints, yields a more interpretable model than the standard deep learning model. The dictionary learned by sparse coding can be more easily understood and the activations of these elements creates a selective feature output.
  arXiv  Detail & Related papers  (2020-11-24T00:26:40Z)
• MetaSDF: Meta-learning Signed Distance Functions [85.81290552559817]
  Generalizing across shapes with neural implicit representations amounts to learning priors over the respective function space. We formalize learning of a shape space as a meta-learning problem and leverage gradient-based meta-learning algorithms to solve this task.
  arXiv  Detail & Related papers  (2020-06-17T05:14:53Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.