Cover Image

Automated Interpretability Metrics Do Not Distinguish Trained and Random Transformers

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Heap, Thomas, Lawson, Tim, Farnik, Lucy, Aitchison, Laurence
Format:	Preprint
Published:	2025
Subjects:	Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Residual Stream Analysis with Multi-Layer SAEs
by: Lawson, Tim, et al.
Published: (2024)

Jacobian Sparse Autoencoders: Sparsify Computations, Not Just Activations
by: Farnik, Lucy, et al.
Published: (2025)

Learning to Skip the Middle Layers of Transformers
by: Lawson, Tim, et al.
Published: (2025)

Massively Parallel Expectation Maximization For Approximate Posteriors
by: Heap, Thomas, et al.
Published: (2025)

Why you don't overfit, and don't need Bayes if you only train for one epoch
by: Aitchison, Laurence
Published: (2024)

Controlling changes to attention logits
by: Anson, Ben, et al.
Published: (2025)

Batch size invariant Adam
by: Wang, Xi, et al.
Published: (2024)

How to set AdamW's weight decay as you scale model and dataset size
by: Wang, Xi, et al.
Published: (2024)

Using Autodiff to Estimate Posterior Moments, Marginals and Samples
by: Bowyer, Sam, et al.
Published: (2023)

Function-Space Learning Rates
by: Milsom, Edward, et al.
Published: (2025)

Using Neural Networks for Data Cleaning in Weather Datasets
by: Hanslope, Jack R. P., et al.
Published: (2024)

Flexible Infinite-Width Graph Convolutional Neural Networks
by: Anson, Ben, et al.
Published: (2024)

Convolutional Deep Kernel Machines
by: Milsom, Edward, et al.
Published: (2023)

Stochastic Kernel Regularisation Improves Generalisation in Deep Kernel Machines
by: Milsom, Edward, et al.
Published: (2024)

Scale-invariant Attention
by: Anson, Ben, et al.
Published: (2025)

MONGOOSE: Path-wise Smooth Bayesian Optimisation via Meta-learning
by: Yang, Adam X., et al.
Published: (2023)

Position: Don't Use the CLT in LLM Evals With Fewer Than a Few Hundred Datapoints
by: Bowyer, Sam, et al.
Published: (2025)

Bayesian Low-rank Adaptation for Large Language Models
by: Yang, Adam X., et al.
Published: (2023)

Interpretable Clustering with the Distinguishability Criterion
by: Turfah, Ali, et al.
Published: (2024)

Inverse-Free Sparse Variational Gaussian Processes
by: Cortinovis, Stefano, et al.
Published: (2026)

Rodent-Bench
by: Heap, Thomas, et al.
Published: (2026)

STARC: A General Framework For Quantifying Differences Between Reward Functions
by: Skalse, Joar, et al.
Published: (2023)

Leveraging Interpretability in the Transformer to Automate the Proactive Scaling of Cloud Resources
by: Ba, Amadou, et al.
Published: (2024)

Learning Generation Orders for Masked Discrete Diffusion Models via Variational Inference
by: Fox, David, et al.
Published: (2026)

Bayesian Reward Models for LLM Alignment
by: Yang, Adam X., et al.
Published: (2024)

Questionable practices in machine learning
by: Leech, Gavin, et al.
Published: (2024)

Graph Homomorphism Distortion: A Metric to Distinguish Them All and in the Latent Space Bind Them
by: Carrasco, Martin, et al.
Published: (2025)

Prior Aware Memorization: An Efficient Metric for Distinguishing Memorization from Generalization in Large Language Models
by: Tiwari, Trishita, et al.
Published: (2026)

Inducing Human-like Biases in Moral Reasoning Language Models
by: Karpov, Artem, et al.
Published: (2024)

Surrogate Fitness Metrics for Interpretable Reinforcement Learning
by: Altmann, Philipp, et al.
Published: (2025)

Symmetry-Aware Transformer Training for Automated Planning
by: Fritzsche, Markus, et al.
Published: (2025)

Structured Transformations for Stable and Interpretable Neural Computation
by: Nikooroo, Saleh, et al.
Published: (2025)

Machine learning emulation of precipitation from km-scale UK regional climate simulations using a diffusion model
by: Addison, Henry, et al.
Published: (2024)

Metric as Transform: Exploring beyond Affine Transform for Interpretable Neural Network
by: Sapkota, Suman
Published: (2024)

Symmetry Breaking in Transformers for Efficient and Interpretable Training
by: Silverstein, Eva, et al.
Published: (2026)

Random-Effects Algorithm for Random Objects in Metric Spaces
by: Matabuena, Marcos, et al.
Published: (2026)

When Are Two Networks the Same? Tensor Similarity for Mechanistic Interpretability
by: Gonzalez, ML Nissen, et al.
Published: (2026)

Automated Classification of Volcanic Earthquakes Using Transformer Encoders: Insights into Data Quality and Model Interpretability
by: Suzuki, Y., et al.
Published: (2025)

Deterministic Bounds and Random Estimates of Metric Tensors on Neuromanifolds
by: Sun, Ke
Published: (2025)

ML Interpretability: Simple Isn't Easy
by: Räz, Tim
Published: (2022)