cd/entity/SAEBench· home› entities› SAEBench
grep -l @saebench /news/*.json | wc -l → 1

SAEBench

mentions 1 type Organization feed RSS

// recent coverage 1 mentions

04:07
2026-10-07
arxiv.org
ai-safety

Backdooring Sparse Autoencoders

A paper submitted to arXiv on 5 October 2026 introduces a decoder-only sparse autoencoder (SAE) backdoor that induces attacker-chosen behavior when the modified SAE is inserted into the forward pass o…

// co-occurs with top 3 entities
arXiv 1 Sparse autoencoders 1 HumanEval 1
// topics top 5 topics
ai safety 1 ai research 1 large language models 1 machine learning 1 artificial intelligence 1