Vrinda Kohli

Vrinda Kohli

OS-HARM: The AI Safety Benchmark That Puts LLM Agents Through Hell

OS-HARM: The AI Safety Benchmark That Puts LLM Agents Through Hell

Language models have come a long way. From playing autocomplete in your email to writing decent Python scripts, they’ve now levelled up into agents: full-blown task-doers who can click, scroll, type, and wreak havoc across your desktop. These “computer use agents” are smart enough to open your emails, edit

Your Horrible Code is Making LLMs Evil: Exploring Emergent Misalignment

Your Horrible Code is Making LLMs Evil: Exploring Emergent Misalignment

What is Emergent Misalignment? One bad apple can spoil the bunch. Apparently this stands true when speaking of finetuning tasks too. A recent paper uncovered a quite interesting phenomenon: finetuning an LLM on insecure code led it to show homicidal tendencies in conversations. And this is not just a fluke,

Sure your LLM is smart, but does it really give a damn?

Sure your LLM is smart, but does it really give a damn?

You can take your model to the water, but you can’t make it think. Every frontier lab’s model drops are accompanied by boasts on improved capabilities on a dozen benchmarks. A recent study explores that the fact that a model is capable of accomplishing a task doesn’t

Making Language Models Unbiased, One Vector At a Time

Making Language Models Unbiased, One Vector At a Time

Introduction AI has officially broken out of the tech bubble and into everyday workflows, boosting productivity but also raising safety concerns, especially around bias in large language models. These models inherit societal biases from internet data, and debiasing efforts by frontier labs can sometimes go too far (remember the racially