Personalize-then-Store : Benchmarking and Learning Personalized Memory for Long-horizon Agents
Yeonjun In, Wonjoong Kim, Sangwu Park, Kanghoon Yoon, Chanyoung Park · 2026
Read Personalize-then-Store : Benchmarking and Learning Personalized Memory for Long-horizon Agents by Yeonjun In, Wonjoong Kim, Sangwu Park, Kanghoon Yoon, Chanyoung Park (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-2605-25535.txt.
Description
Yeonjun In and colleagues ask whether language-model memory should keep the same items for every user. They argue that one fixed rule wastes a tight memory budget on passing chats while losing context that matters over long tasks. Their PerMemBench testbed spans multi-year, multi-domain histories for varied personas, and a session-level gating scheme skips storage for transient sessions. Personalization lifts retention when gating is perfect, yet reliable gating itself stays unsolved.
Catalog record
| CREATOR | Yeonjun In, Wonjoong Kim, Sangwu Park, Kanghoon Yoon, Chanyoung Park |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 83471 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.25535