Checked fact 11826 Sep 2026Research

The abstract states: 'Adaptive attacks using score feedback double the mean highest attacker-target probability found during optimization, while success on fresh validation calls rises from 1.8% to 3.5%.'

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

Adaptive attacks using score feedback double the mean highest attacker-target probability found during optimization, while success on fresh validation calls rises from 1.8% to 3.5%.

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

The source

Decision Hijacking: Prompt Injection Attacks on Jev Typed Probabilistic Decisions
arXiv · 2026-09-23

Open the source archived copy

Checked

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

In the story

Paper finds prompt injection can shift Jev's typed decisions, though rarely to the attacker's target 26 Sep 2026

Cite this fact

Short permanent address

In a sentence

“The abstract states: 'Adaptive attacks using score feedback double the mean highest attacker-target probability found during optimization, while success on fresh validation calls rises from 1.8% to 3.5%.'.” ai notis, checked 26 Sep 2026. Source: Decision Hijacking: Prompt Injection Attacks on Jev Typed Probabilistic Decisions (arXiv). https://ainotis.com/c/118

As a reference (APA style)

ai notis. (2026, September 26). The abstract states: 'Adaptive attacks using score feedback double the mean highest attacker-target probability found during optimization, while success on fresh validation calls rises from 1.8% to 3.5%.' [Checked claim 118]. ai notis. https://ainotis.com/c/118

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