RPS: Information Elicitation with Reinforcement Prompt Selection

TL;DR

This paper introduces RPS, a reinforcement learning framework for adaptive prompt selection to improve information elicitation in open-ended dialogues with large language models.

Contribution

It proposes a novel RL-based approach for prompt selection and introduces IELegal, a new dataset for evaluating dialogue-based information elicitation.

Findings

RL agent outperforms random query baseline in synthetic experiments.

RPS outperforms static prompt baselines on the IELegal dataset.

Adaptive prompt selection enhances information elicitation in dialogue systems.

Abstract

Large language models (LLMs) have shown remarkable capabilities in dialogue generation and reasoning, yet their effectiveness in eliciting user-known but concealed information in open-ended conversations remains limited. In many interactive AI applications, such as personal assistants, tutoring systems, and legal or clinical support, users often withhold sensitive or uncertain information due to privacy concerns, ambiguity, or social hesitation. This makes it challenging for LLMs to gather complete and contextually relevant inputs. In this work, we define the problem of information elicitation in open-ended dialogue settings and propose Reinforcement Prompt Selection (RPS), a lightweight reinforcement learning framework that formulates prompt selection as a sequential decision-making problem. To analyze this problem in a controlled setting, we design a synthetic experiment, where a reinforcement learning agent outperforms a random query baseline, illustrating the potential of policy-based approaches for adaptive information elicitation. Building on this insight, RPS learns a policy over a pool of prompts to adaptively elicit concealed or incompletely expressed information from users through dialogue. We also introduce IELegal, a new benchmark dataset constructed from real legal case documents, which simulates dialogue-based information elicitation tasks aimed at uncovering case-relevant facts. In this setting, RPS outperforms static prompt baselines, demonstrating the effectiveness of adaptive prompt selection for eliciting critical information in LLM-driven dialogue systems.