Token-permutation obfuscation preserves GPT-2-small's performance on an indirect-object task but makes head-level logit attribution and prompt reconstruction noisier, according to a from-scratch experiment with an uncontrolled baseline.
The Urgency of Interpretability
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CR 1years
2025 1verdicts
REJECT 1roles
background 1polarities
support 1representative citing papers
citing papers explorer
-
Mechanistic Interpretability in the Presence of Architectural Obfuscation
Token-permutation obfuscation preserves GPT-2-small's performance on an indirect-object task but makes head-level logit attribution and prompt reconstruction noisier, according to a from-scratch experiment with an uncontrolled baseline.