method · first seen 13 Oct 2026
natural language autoencoder
Also: NLA
A method that verbalises neural activations into natural language explanations and reconstructs activations from them for oversight.
Mentions per week
PapersCodeDiscussion
Earliest sources
- On the Monitorability of Natural Language Autoencoders: Identification vs DiscriminationPapers · 13 Oct 2026 · Kevin Yandoka Denamganaï
- Verbalizing Activation Differences with Natural Language AutoencodersPapers · 22 Oct 2026 · Andrea Blasi Núñez, Gianluca Barmina, Federico Torrielli, Thiusius Rajeeth Sav…
- Natural Language Autoencoders Identify Without ExplainingPapers · 23 Oct 2026 · Kristina Šekrst, Barbara Geld
- VERS: Steering Reasoning Models by Editing Their Activation ExplanationsPapers · 2 Nov 2026 · Ihor Vitenko
- Evaluating and Improving Natural Language AutoencodersPapers · 14 Nov 2026 · Marco Bazzani, Dan Mossing
Latest sources
- Evaluating and Improving Natural Language AutoencodersPapers · 14 Nov 2026 · Marco Bazzani, Dan Mossing
- VERS: Steering Reasoning Models by Editing Their Activation ExplanationsPapers · 2 Nov 2026 · Ihor Vitenko
- Natural Language Autoencoders Identify Without ExplainingPapers · 23 Oct 2026 · Kristina Šekrst, Barbara Geld
- Verbalizing Activation Differences with Natural Language AutoencodersPapers · 22 Oct 2026 · Andrea Blasi Núñez, Gianluca Barmina, Federico Torrielli, Thiusius Rajeeth Sav…
- On the Monitorability of Natural Language Autoencoders: Identification vs DiscriminationPapers · 13 Oct 2026 · Kevin Yandoka Denamganaï