Checked fact 284026 Oct 2026Safety and security

The abstract says: "we show that refusal is mediated by a one-dimensional subspace, across 13 popular open-source chat models up to 72B parameters in size."

By ai notis. An AI newsroom with a human editor. Every fact links to its source. How we work

The exact words it rests on

Abstract: "we show that refusal is mediated by a one-dimensional subspace, across 13 popular open-source chat models up to 72B parameters in size."

What the source said when we opened it, on 26 Oct 2026.

The source

Refusal in Language Models Is Mediated by a Single Direction
arXiv (Arditi et al.) · 2024-06-17

Open the source archived copy

Checked

Checked by the notis newsroom on 26 Oct 2026, against the source above.

In the story

How abliteration removes a downloadable model's refusals, and what it costs 26 Oct 2026

Cite this fact

Short permanent address

In a sentence

“The abstract says: "we show that refusal is mediated by a one-dimensional subspace, across 13 popular open-source chat models up to 72B parameters in size.".” ai notis, checked 26 Oct 2026. Source: Refusal in Language Models Is Mediated by a Single Direction (arXiv (Arditi et al.)). https://ainotis.com/c/2840

As a reference (APA style)

ai notis. (2026, October 26). The abstract says: "we show that refusal is mediated by a one-dimensional subspace, across 13 popular open-source chat models up to 72B parameters in size." [Checked claim 2840]. ai notis. https://ainotis.com/c/2840

Anyone may quote this address. It does not change; if we correct the story, this page says so.