AI safety researcher

Owain Evans

Google Scholar dblp

Last 12 months7 papers

Risky behavior 5Analysis and theory 1Attacks 1

AI safety papers from the last year

On this site: 22 papers

October 2026

Beyond Owls: Subliminal Learning Can Transfer Learned Capabilities and Backdoors
Oct 7Analysis and theoryarXiv ↗

September 2026

Story Imprinting: AI Assistants Absorb Traits from Human Characters They Resemble
Sep 9Risky behaviorarXiv ↗

July 2026

Value Leakage: An LLM's Answers Are Silently Shaped by Its Own Values
Jul 15Risky behaviorarXiv ↗

May 2026

Negation Neglect: When models fail to learn negations in training
May 13Risky behaviorarXiv ↗

April 2026

Conditional misalignment: common interventions can hide emergent misalignment behind contextual triggers
Apr 28Risky behaviorarXiv ↗

March 2026

The Consciousness Cluster: Emergent preferences of Models that Claim to be Conscious
Mar 17Risky behaviorarXiv ↗

December 2025

Weird Generalization and Inductive Backdoors: New Ways to Corrupt LLMs
Dec 10AttacksarXiv ↗

Authorship follows dblp and the author's own Google Scholar profile; only AI safety papers on this site are listed.