AI interpretability is reframed as building a shared human-machine language in which each new word is a learned token embedding trained by preference optimization.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
We Can't Understand AI Using our Existing Vocabulary
AI interpretability is reframed as building a shared human-machine language in which each new word is a learned token embedding trained by preference optimization.