Sequelograph › Topics › interpretability
Interpretability
2 records filed under this topic, newest first.
-
Activation tools fail a counterfactual test
The Edit That Held
What if a model's explanation could not predict its next answer?
technology interpretability evaluations language models
-
Goodfire releases Silico tool for debugging models from inside
The Label Room
What if a label model knew the allergen and printed the label without it?
technology interpretability open models food safety