A training-free attribution method that estimates the relative influence of upstream neurons and attention outputs on a target neuron's activation and identifies candidate sparse pathways with potential implications for efficient inference is introduced.
Abstract
Feedforward network (FFN) blocks account for a large fraction of the parameters and computation in Transformer architectures, yet their internal structure remains difficult to interpret due to the additive superposition induced by the residual stream. We examine whether the activation of an FFN neuron can be explained by a sparse set of preceding neuron activations and attention outputs. We introduce a training-free attribution method that estimates the relative influence of upstream neurons and attention outputs on a target neuron's activation. Empirically, across models and layers, we find that small subsets of preceding activations and attention outputs suffice to preserve neuron activations with high fidelity when all remaining inputs are masked with their average values. Effective sparsity is even greater when accounting for the inherent activation sparsity of upstream layers. Moreover, applying the neuron-specific masks in all layers simultaneously, such that the induced deviations propagate through the network, leaves model perplexity largely unchanged at moderate sparsity levels. These results demonstrate that, despite dense parameterization, FFNs exhibit sparse and structured inter-layer dependencies at the neuron level. Our method provides a practical, scalable tool for circuit-level interpretability and identifies candidate sparse pathways with potential implications for efficient inference.
A von-Neumann inspired hypothesis of efficient computation as an alternative for neural decoding, a memory-augmented Transformer whose feed-forward block is a low-rank instruction bank: a shared base operator plus a small set of learned low-rank instructions, from which a per-token code synthesizes the weight matrix actually used at that token.
Superposition refers to neural networks representing more features than they have dimensions. It offers a possible explanation for polysemantic neurons and motivates methods for recovering interpretable features from neural activations. Theoretical models typically start with a given set of input features and assumptions about how their values vary across inputs, then study how a network encodes those values in a lower-dimensional hidden representation. Empirical work, by contrast, seeks to identify the features encoded in trained networks and determine their role in computation. In this survey, we review the geometry, learning, and computation of superposed representations, explaining how feature statistics and decoder choice affect the conclusions. To connect these theoretical accounts with evidence from trained networks, we compare practical methods for recovering and analyzing features and examine what their evaluations establish. Since accurate activation reconstruction alone does not establish feature identity or causal use, we discuss the methods'documented failures and applications in light of the evidence available for these different claims. Finally, we assess previously stated open problems and identify remaining theoretical and empirical questions about superposition in trained networks. We hope our work can pave the way for a deeper understanding of superposition and more reliable methods for interpreting neural networks.
Prox is a two-stage training-free framework for sparse SwiGLU FFNs that outperforms training-free baselines at all sparsity levels, achieves up to a $1.99\times end-to-end decoding speedup at 70\% FFN sparsity, and is compatible with quantization and sparse attention.
A family of difference-informed pruning methods built upon this principle are introduced, suggesting that preserving output differences is a broadly useful and composable signal for post-training LLM sparsification.
Linghao Kong, Inimai Subramanian, Micah Adler et al.· 0 citations
This work introduces a method that maintains near-baseline accuracy, induces usage-based modularity by sparsely routing inputs to neuron groups, and encourages specialization of these modules, such that their activations are correlated with input classes.
Overparameterized neural networks carry far more hidden units than a task nominally requires, raising the question of which neurons are essential and whether that distinction is legible in the representation itself, without labels or gradients. We cast neuron selection as the problem of coarse-graining the hidden layer by retaining a subset of its neurons, and score each putative selection by the mapping entropy (ME). This quantity measures the loss of discriminatory power inherent in discarding part of the network neurons, and the selection that minimises the ME is taken as particularly informative. This criterion is fully unsupervised, in that it depends only on hidden-activation statistics. In teacher-student networks, ME optimisation recovers the minimal teacher-consistent representation and retains extra units in proportion to the hidden layer's residual variability; in a non-linear Gaussian process task, it selects coherent functional-class mappings whose preferred class shifts across training. On this task and on translation-augmented MNIST, ME-selected subnetworks outperform random subsets of equal size, most clearly under strong compression - linking configurational distinguishability to predictive performance.
Margherita Mele, Andrea Castagna, Roberto Menichetti et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
