Pith. sign in

REVIEW 9 cited by

Bilinear MLPs enable weight-based mechanistic interpretability

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2410.08417 v2 pith:BC3FLE7L submitted 2024-10-10 cs.LG stat.ML

Bilinear MLPs enable weight-based mechanistic interpretability

Michael T. Pearce , Thomas Dooms , Alice Rigg , Jose M. Oramas , Lee Sharkey This is my paper
classification cs.LG stat.ML
keywords bilinearmlpsweightsinterpretabilityunderstandingcurrentelement-wisefeatures
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

A mechanistic understanding of how MLPs do computation in deep neural networks remains elusive. Current interpretability work can extract features from hidden activations over an input dataset but generally cannot explain how MLP weights construct features. One challenge is that element-wise nonlinearities introduce higher-order interactions and make it difficult to trace computations through the MLP layer. In this paper, we analyze bilinear MLPs, a type of Gated Linear Unit (GLU) without any element-wise nonlinearity that nevertheless achieves competitive performance. Bilinear MLPs can be fully expressed in terms of linear operations using a third-order tensor, allowing flexible analysis of the weights. Analyzing the spectra of bilinear MLP weights using eigendecomposition reveals interpretable low-rank structure across toy tasks, image classification, and language modeling. We use this understanding to craft adversarial examples, uncover overfitting, and identify small language model circuits directly from the weights alone. Our results demonstrate that bilinear layers serve as an interpretable drop-in replacement for current activation functions and that weight-based interpretability is viable for understanding deep-learning models.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 9 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. WriteSAE: Sparse Autoencoders for Recurrent State

    cs.LG 2026-05 unverdicted novelty 8.0

    WriteSAE introduces sparse autoencoders with rank-1 matrix atoms for recurrent state updates, allowing replacement tests that outperform deletion on 92.4% of positions and a formula predicting logit changes with R²=0.98.

  2. WriteSAE: Sparse Autoencoders for Recurrent State

    cs.LG 2026-05 unverdicted novelty 8.0

    WriteSAE is the first sparse autoencoder that factors decoder atoms into the native d_k x d_v cache write shape of recurrent models and supplies a closed-form per-token logit shift for atom substitution.

  3. WriteSAE: Sparse Autoencoders for Recurrent State

    cs.LG 2026-05 unverdicted novelty 8.0

    WriteSAE decomposes recurrent model cache writes into substitutable atoms with a closed-form logit shift, achieving high substitution success and targeted behavioral installs on models like Qwen3.5 and Mamba-2.

  4. When Are Two Networks the Same? Tensor Similarity for Mechanistic Interpretability

    cs.LG 2026-05 unverdicted novelty 7.0

    Tensor similarity is a symmetry-invariant metric that measures functional equivalence between tensor-based networks using a recursive algorithm for cross-layer mechanisms.

  5. WriteSAE: Sparse Autoencoders for Recurrent State

    cs.LG 2026-05 unverdicted novelty 7.0

    WriteSAE factors sparse autoencoder decoder atoms to the native d_k x d_v cache write shape in recurrent models, provides a closed-form logit shift, and demonstrates high success in atom substitution and behavioral ed...

  6. Task complexity shapes internal representations and robustness in neural networks

    cs.LG 2025-08 unverdicted novelty 7.0

    Harder classification tasks produce neural representations whose accuracy collapses under binarization and shuffling while easier tasks remain robust, defining task complexity via the performance gap between full-prec...

  7. Individual Parameters in Weight-Sparse Transformers Appear Interpretable

    cs.LG 2026-07 conditional novelty 6.5

    An automated LLM pipeline finds that 12–31% of nonzero weights in weight-sparse transformers admit short, held-out-validated descriptions of when they matter, far above dense controls.

  8. Legible-by-Construction: Attention and End-to-End Transformers

    cs.CL 2026-07 conditional novelty 6.0

    Bounding attention values into fuzzy memberships or Boolean set operations yields 44–62% crisp selective detectors and an end-to-end legible transformer at baseline quality.

  9. Preserving Bilinear Weight Spectra with a Signed and Shrunk Quadratic Activation Function

    cs.LG 2025-09 reject novelty 5.0

    SQS is a new GLU activation that reportedly preserves bilinear weight spectra for interpretability while achieving competitive performance on MNIST, FMNIST, and Tiny Stories.