Skip to content
onsuggest
Log in
Article en

False faced: can a simple classifier tell when a model is manipulating its activations to evade probes?

Source: LessWrong ·

This experiment was a submission to my Statistics & Probability course Bayesian classifiers project. TL;DR: We test whether simple classifiers can tell when a model is hiding concepts in its activations to evade probes. We find that, for Bayesian classifiers, the probabilities are enough to separate most evading rows from the rest on concepts seen in training, but the correlation aspect through the pairs of dimensions that the Dependent version adds over the Naive approach does not measurably help. None of the classifiers work on a new concept, because they learn the trigger word that elicits

What is False faced: can a simple classifier tell when a model is manipulating its activations to evade probes??

False faced: can a simple classifier tell when a model is manipulating its activations to evade probes? is a article discovery curated from LessWrong. This experiment was a submission to my Statistics & Probability course Bayesian classifiers project. TL;DR: We test whether simple classifiers can tell when a model is hiding…

How can I access or experience this?

You can experience this content directly on its original host (LessWrong) by clicking the "Open Original" button.

Where does this content come from?

Curated from LessWrong by the OnSuggest team to surface the most interesting and serendipitous content on the web.

Keyboard shortcuts

Previous discovery
←P
Next discovery
→N
Like
L
Not for me
X
Save
S
Show this help
?
Close
Esc

Shortcuts work on the discovery page.