Deliberative Searcher: Improving LLM Reliability via Reinforcement Learning with constraints
Not on the map. Why?
Abstract (arXiv)
Improving the reliability of large language models (LLMs) is critical for deploying them in real-world scenarios. In this paper, we propose Deliberative Searcher, the first framework to integrate certainty calibration with retrieval-based search for open-domain question answering. The agent performs multi-step reflection and verification over Wikipedia data and is trained with a reinforcement learning algorithm that optimizes for accuracy under a soft reliability constraint. Empirical results show that proposed method improves alignment between model confidence and correctness, leading to more trustworthy outputs. This paper will be continuously updated.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 24 of its 56 references that are in the map. First posted on arXiv 22 Jul 2025.
Builds on
24 of 56 references.
References · 24
- 2020Language Models are Few-Shot Learnerscited by 32,419
- 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
- 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
- 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
- 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,626
- 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,754
- 2023InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Taskscited by 2,565
- 2021WebGPT: Browser-assisted question-answering with human feedbackcited by 1,745
- 2025Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learningcited by 1,444
- 2020Constructing A Multi-hop QA Dataset for Comprehensive Evaluation of Reasoning Stepscited by 1,282
- 2017Constrained Policy Optimizationcited by 983
- 2023Just Ask for Calibration: Strategies for Eliciting Calibrated Confidence Scores from Language Models Fine-Tuned with Human Feedbackcited by 812
12 more
- 2022Teaching Models to Express Their Uncertainty in Wordscited by 751
- 2023Safe RLHF: Safe Reinforcement Learning from Human Feedbackcited by 656
- 2018Reward Constrained Policy Optimizationcited by 417
- 2025xbench: Tracking Agents Productivity Scaling with Profession-Aligned Real-World Evaluationscited by 120
- 2023Let's Sample Step by Step: Adaptive-Consistency for Efficient Reasoning and Coding with LLMscited by 115
- 2024Escape Sky-high Cost: Early-stopping Self-Consistency for Multi-step Reasoningcited by 108
- 2024LongRAG: Enhancing Retrieval-Augmented Generation with Long-context LLMscited by 108
- 2024SaySelf: Teaching LLMs to Express Confidence with Self-Reflective Rationalescited by 105
- 2019Safe Policies for Reinforcement Learning via Primal-Dual Methodscited by 79
- 2024KnowAgent: Knowledge-Augmented Planning for LLM-Based Agentscited by 61
- 2025Reasoning about Uncertainty: Do Reasoning Models Know When They Don't Know?cited by 38
- 2025GAIA: A Global, Multi-modal, Multi-scale Vision-Language Dataset for Remote Sensing Image Analysiscited by 7
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2025Do LLM Agents Know How to Ground, Recover, and Assess? A Benchmark for Epistemic Competence in Information-Seeking Agents
- 2025YESciEval: Robust LLM-as-a-Judge for Scientific Question Answering
- 2024BIDER: Bridging Knowledge Inconsistency for Efficient Retrieval-Augmented LLMs via Key Supporting Evidence
- 2025Learning the Boundary of Solvability: Aligning LLMs to Detect Unsolvable Problems
- 2026Confidence Before Answering: A Paradigm Shift for Efficient LLM Uncertainty Estimation
- 2025Rewarding Doubt: A Reinforcement Learning Approach to Calibrated Confidence Expression of Large Language Models
- 2025O2-Searcher: A Searching-based Agent Model for Open-Domain Open-Ended Question Answering
- 2025Search Wisely: Mitigating Sub-optimal Agentic Searches By Reducing Uncertainty