CUPID: Evaluating Personalized and Contextualized Alignment of LLMs from Interactions

TL;DR

CUPID is a benchmark designed to evaluate how well large language models can infer and adapt to users' dynamic preferences based on multi-turn interaction histories, highlighting current limitations in contextual personalization.

Contribution

This work introduces CUPID, a new benchmark with 756 interaction sessions to assess LLMs' ability to infer user preferences from context and interactions.

Findings

State-of-the-art LLMs have under 50% precision in inferring preferences.

LLMs struggle to identify relevant past context for new requests.

Current models achieve only 65% recall in preference inference.

Abstract

Personalization of Large Language Models (LLMs) often assumes users hold static preferences that reflect globally in all tasks. In reality, humans hold dynamic preferences that change depending on the context. As users interact with an LLM in various contexts, they naturally reveal their contextual preferences, which a model must infer and apply in future contexts to ensure alignment. To assess this, we introduce CUPID, a benchmark of 756 human-curated interaction session histories between users and LLM-based chat assistants. In each interaction session, the user provides a request in a specific context and expresses their preference through multi-turn feedback. Given a new user request and prior interaction sessions, our benchmark assesses whether LLMs can infer the preference relevant to this request and generate a response that satisfies this preference. With CUPID, we evaluated 10 open and proprietary LLMs, revealing that state-of-the-art LLMs struggle to infer preferences from multi-turn interactions and fail to discern what previous context is relevant to a new request -- under 50% precision and 65% recall. Our work highlights the need to advance LLM capabilities for more contextually personalized interactions and proposes CUPID as a resource to drive these improvements.