# Closing the Expression Gap in LLM Instructions via Socratic Questioning

arXiv: https://arxiv.org/abs/2510.27410 · HTML: https://mlsanity.com/paper/2510.27410/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 40 of its 62 references that are in the map.
- First posted on arXiv 31 Oct 2025.

## Builds on (most cited first, 30 of 40)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Deep reinforcement learning from human preferences](https://mlsanity.com/paper/1706.03741/) (2017)
- [Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problems](https://mlsanity.com/paper/2005.01643/) (2020)
- [Offline Reinforcement Learning with Implicit Q-Learning](https://mlsanity.com/paper/2110.06169/) (2021)
- [MultiWOZ -- A Large-Scale Multi-Domain Wizard-of-Oz Dataset for Task-Oriented Dialogue Modelling](https://mlsanity.com/paper/1810.00278/) (2018)
- [RLAIF vs. RLHF: Scaling Reinforcement Learning from Human Feedback with AI Feedback](https://mlsanity.com/paper/2309.00267/) (2023)
- [A Network-based End-to-End Trainable Task-oriented Dialogue System](https://mlsanity.com/paper/1604.04562/) (2016)
- [MultiWOZ 2.2 : A Dialogue Dataset with Additional Annotation Corrections and State Tracking Baselines](https://mlsanity.com/paper/2007.12720/) (2020)
- [Proxy Tasks and Subjective Measures Can Be Misleading in Evaluating Explainable AI Systems](https://mlsanity.com/paper/2001.08298/) (2020)
- [SEED-GRPO: Semantic Entropy Enhanced GRPO for Uncertainty-Aware Policy Optimization](https://mlsanity.com/paper/2505.12346/) (2025)
- [Modeling Future Conversation Turns to Teach LLMs to Ask Clarifying Questions](https://mlsanity.com/paper/2410.13788/) (2024)
- [Reincarnating Reinforcement Learning: Reusing Prior Computation to Accelerate Progress](https://mlsanity.com/paper/2206.01626/) (2022)
- [Prompting Large Language Models With the Socratic Method](https://mlsanity.com/paper/2303.08769/) (2023)
- [Uncertainty of Thoughts: Uncertainty-Aware Planning Enhances Information Seeking in Large Language Models](https://mlsanity.com/paper/2402.03271/) (2024)
- [Guided Dialog Policy Learning: Reward Estimation for Multi-Domain Task-Oriented Dialog](https://mlsanity.com/paper/1908.10719/) (2019)
- [Large-Scale Multi-Domain Belief Tracking with Knowledge Sharing](https://mlsanity.com/paper/1807.06517/) (2018)
- [Active Preference Inference using Language Models and Probabilistic Reasoning](https://mlsanity.com/paper/2312.12009/) (2023)
- [Asking Clarification Questions to Handle Ambiguity in Open-Domain QA](https://mlsanity.com/paper/2305.13808/) (2023)
- [Teaching Language Models to Evolve with Users: Dynamic Profile Modeling for Personalized Alignment](https://mlsanity.com/paper/2505.15456/) (2025)
- [Learning to Ask Informative Questions: Enhancing LLMs with Preference Optimization and Expected Information Gain](https://mlsanity.com/paper/2406.17453/) (2024)
- [A Conditional Variational Framework for Dialog Generation](https://mlsanity.com/paper/1705.00316/) (2017)
- [An Asymptotic Test for Conditional Independence using Analytic Kernel Embeddings](https://mlsanity.com/paper/2110.14868/) (2021)
- [CoCoA: Confidence and Context-Aware Adaptive Decoding for Resolving Knowledge Conflicts in Large Language Models](https://mlsanity.com/paper/2508.17670/) (2025)
- [Dialog Policy Learning for Joint Clarification and Active Learning Queries](https://mlsanity.com/paper/2006.05456/) (2020)
- [Collaborative human-AI trust (CHAI-T): A process framework for active management of trust in human-AI collaboration](https://mlsanity.com/paper/2404.01615/) (2024)
- [Conditional independence testing under misspecified inductive biases](https://mlsanity.com/paper/2307.02520/) (2023)
- [InfoPO: On Mutual Information Maximization for Large Language Model Alignment](https://mlsanity.com/paper/2505.08507/) (2025)

## Abstract (arXiv)

A fundamental bottleneck in human-AI collaboration is the ``intention expression gap," the difficulty for humans to effectively convey complex, high-dimensional thoughts to AI. This challenge often traps users in inefficient trial-and-error loops and is exacerbated by the diverse expertise levels of users. We reframe this problem from passive instruction following to a Socratic collaboration paradigm, proposing an agent that actively probes for information to resolve its uncertainty about user intent. we name the proposed agent Nous, trained to acquire proficiency in this inquiry policy. The core mechanism of Nous is a training framework grounded in the first principles of information theory. Within this framework, we define the information gain from dialogue as an intrinsic reward signal, which is fundamentally equivalent to the reduction of Shannon entropy over a structured task space. This reward design enables us to avoid reliance on costly human preference annotations or external reward models. To validate our framework, we develop an automated simulation pipeline to generate a large-scale, preference-based dataset for the challenging task of scientific diagram generation. Comprehensive experiments, including ablations, subjective and objective evaluations, and tests across user expertise levels, demonstrate the effectiveness of our proposed framework. Nous achieves leading efficiency and output quality, while remaining robust to varying user expertise. In conclusion, our research provides a systematic methodology and a new perspective for addressing the issue of ambiguous intentions in complex human-machine collaboration.