Article

Improving Variant Effect Prediction by Steering Sparse Mechanistic Features in Protein Language Models

2026-05-15

Abstract excerpt

Protein language models (PLMs) like the ESM series encapsulate immense evolutionary knowledge within their high-dimensional continuous embeddings. However, these latent representations are densely entangled, obscuring the fine-grained biophysical constraints necessary for precise functional resolution. To unlock the full expressive power of these embeddings, we propose PLM-SAE, a mechanistic framework that employs...

Topics

Open a Topic to create a Post that cites this publication.

Identifiers and source

Literature Corpus work: 87a6bab4-4517-512e-9759-65b99ca8db03
DOI: 10.64898/2026.05.12.724472

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

Improving Variant Effect Prediction by Steering Sparse Mechanistic Features in Protein Language ModelsDOI 10.64898/2026.05.12.724472Read

Structure-informed protein language models are robust predictors for variant effects.Semantic proximity Read

Protein language models trained on biophysical dynamics inform mutation effects.Semantic proximity Read

Embeddings from protein language models predict conservation and variant effectsSemantic proximity Read

ESM-Effect: An Effective and Efficient Fine-Tuning Framework towards accurate prediction of Mutation’s Functional EffectSemantic proximity Read

Macon: Enhance Protein Mutation Representation using Contrastive Learning with Effect Prediction on Protein–protein InteractionsSemantic proximity Read

Fine-Tuning Protein Language Models Unlocks the Potential of Underrepresented Viral ProteomesSemantic proximity Read

Expert-guided protein Language Models enable accurate and blazingly fast fitness predictionSemantic proximity Read

Enhancing efficiency of protein language models with minimal wet-lab data through few-shot learning.Semantic proximity Read

Graph attention with structural features improves the generalizability of identifying functional sequences at a protein interfaceSemantic proximity Read

MuLAN: Mutation-driven Light Attention Networks for investigating protein-protein interactions from sequencesSemantic proximity Read

From Likelihood to Fitness: Improving Variant Effect Prediction in Protein and Genome Language ModelsSemantic proximity Read

InterPLM: Discovering Interpretable Features in Protein Language Models via Sparse AutoencodersSemantic proximity Read