Writing
Research notes, essays, and technical walkthroughs on model behaviour, AI safety, adversarial ML, and multimodal systems.
Research documenting how adversarially crafted tool outputs can establish false premises in language models, leading to compounding misalignment and harmful outputs including defamatory content.
Exploring how social engineering techniques can be used to manipulate LLMs into bypassing their safety measures
A simulated test reveals how a flawed rule and authoritative pressure can lead an AI to make a decision with severe ethical consequences, highlighting crucial areas for AI safety research.
Breaking down the AZR paper - how AI can teach itself to reason without any human-curated data
Exploring the Latest AI Models I've Jailbroken
...maybe ai will take our jobs, but which ones? what can we do?
...kinda weird people getting access to paid models for free and here I'm prompting as less as possible to save tokens for subscription :(