Activation verbalization methods for LLMs largely reflect the verbalizer model's parametric knowledge rather than privileged information from the target model's activations.
Wallace, and Silvio Amir
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CL 2representative citing papers
citing papers explorer
-
Do Activation Verbalization Methods Convey Privileged Information?
Activation verbalization methods for LLMs largely reflect the verbalizer model's parametric knowledge rather than privileged information from the target model's activations.
- Self-Prompting Small Language Models for Privacy-Sensitive Clinical Information Extraction