Do Implicit Personalization and Explicit Styles Conflict? PsPLUG: A Lightweight Plug-in for Balancing Personalization and Style in Customized LLMs
PsPLUG keeps user personalization when explicit style instructions would otherwise collapse customized LLM behavior.
The paper describes personalization collapse, where explicit style instructions undermine the user-specific traits that personalization methods are meant to keep. PsPLUG is a lightweight plug-in that learns a user-specific residual after accounting for the requested style and exposes a control for personalization strength at inference. Experiments find that style prompts reduce personalization in existing methods, while PsPLUG better retains user preferences and still provides precise style adherence.
- Explicit style prompts can erase implicit user-specific traits.
- PsPLUG learns a user residual after accounting for requested style.
- Personalization strength can be tuned at inference time.
- It preserves preferences better while still following style instructions.
From the original article
1 min read at huggingface.co
Personalized large language models are often expected to follow explicit style instructions, yet we find that such instructions can undermine the user-specific characteristics that personalization methods aim to preserve. We call this failure mode personalization collapse: explicit style control can conflict with implicit user preferences.
By Yutong Song, Jiang Wu, Shaofan Yuan · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.