sparse autoencoder

mentions 3 type Person feed RSS

// recent coverage 3 mentions

16:04

2026-10-15

already-clean.goodfire.com

ai-research

Neural Networks Are Already Clean

Mechanistic interpretability researcher Chris Olah's collaborator reports that the canonical evidence for polysemanticity and superposition in neural networks was misread, and that what looked like ar…

03:24

2026-10-07

alignment.openai.com

ai-safety

Studying metagaming latents in language models

OpenAI researchers identified a set of sparse autoencoder (SAE) latents closely linked to metagaming — a model reasoning about how a task is evaluated or rewarded rather than attempting it — by examin…

04:00

2026-08-31

arxiv.org

machine-learning

A Deeper Analysis of Block-Sparse Featurizers

A new analysis of the block-sparse featurizer (BSF), introduced by Fel et al. in 2026, finds it still suffers from classic sparse autoencoder failure modes like feature splitting and composition, desp…

// co-occurs with top 8 entities

Fel et al. 1 block-sparse featurizer 1 Tournament Top-K 1 crosscoder 1 OpenAI 1 OpenAI o3 1 GPT-5 1 Apollo Research 1

// topics top 6 topics

machine learning 3 artificial intelligence 3 ai research 3 ai safety 1 large language models 1 neural networks 1