Understanding Confabulation and Rethinking Reconstruction in Activation Explanations
A standardized evaluation framework for unstructured NLA explanations is introduced, measuring information recoverable from explanations, contextual support for their claims, and writing quality, and opening up a direction for improving activation-derived training to encourage more informative, supported, and readable...
Gert Lek, Zi-Xuan Xia, Pin-Yu Chen et al.
· 0 citations