Universal Activation Verbalizer: A Unified Framework for Cross-Model Activation Explanation
Haiyan Zhao, Zirui He, Guanchu Wang, Ali Payani, Yingcong Li Mengnan Du · 2026
Read Universal Activation Verbalizer: A Unified Framework for Cross-Model Activation Explanation by Haiyan Zhao, Zirui He, Guanchu Wang, Ali Payani, Yingcong Li Mengnan Du (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-25903.txt.
Description
Zhao and colleagues present the Universal Activation Verbalizer, showing donor codes with a shared decoder. It learns an adapter turning codes into soft tokens. It also allows adapter-only transfer: a frozen decoder side LoRA stays put while a new adapter is trained. Across classification, retrieval and summarization it matches self explanation, enabling cross-model verbalization. Ablations show decoder side tuning improves behavior, while the adapter supplies grounded information.
Catalog record
| CREATOR | Haiyan Zhao, Zirui He, Guanchu Wang, Ali Payani, Yingcong Li Mengnan Du |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 92055 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.25903