Record
- Authors
- Angsten, T.
- Year
- 2026
- Title
- Adversarial Training Can Further Hide a Backdoor, but Does It Teach the Model to Better Identify Its Trigger
- Published in
- LessWrong, 5 November 2026
- Type
- Web page
- Links
- https://www.lesswrong.com/posts/uKFrPK2oQo8LECrbh
- Our note
- none yet
- Verification
- url-ok: the web address responds (the page itself is not checked).
https://www.lesswrong.com/posts/uKFrPK2oQo8LECrbh: HTTP 200, checked 2026-11-14 - Id
angsten-2026-adversarial-training-can-further-hide-backdoor
Cited on
- Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training read status: not stated
As cited
The pages cite this work in one form.
-
Angsten, T. (2026). Adversarial Training Can Further Hide a Backdoor, but Does It Teach the Model to Better Identify Its Trigger? LessWrong, 5 November 2026. https://www.lesswrong.com/posts/uKFrPK2oQo8LECrbh
BibLaTeX
@online{angsten-2026-adversarial-training-can-further-hide-backdoor,
author = {Angsten, T.},
title = {{Adversarial Training Can Further Hide a Backdoor, but Does It Teach the Model to Better Identify Its Trigger}},
organization = {LessWrong, 5 November 2026},
date = {2026},
url = {https://www.lesswrong.com/posts/uKFrPK2oQo8LECrbh}
}