QueryBandits for Hallucination Mitigation: Exploiting Semantic Features for No-Regret Rewriting

• URL: http://arxiv.org/abs/2508.16697v1
• Date: Fri, 22 Aug 2025 01:41:49 GMT
• Title: QueryBandits for Hallucination Mitigation: Exploiting Semantic Features for No-Regret Rewriting
• Abstract summary: We introduce QueryBandits, a framework that designs rewrite strategies to maximize a reward model.<n>We empirically substantiate the effectiveness of QueryBandits in mitigating hallucination via the intervention that takes the form of a query rewrite.
• Score: 27.117502300924183
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Advanced reasoning capabilities in Large Language Models (LLMs) have caused higher hallucination prevalence; yet most mitigation work focuses on after-the-fact filtering rather than shaping the queries that trigger them. We introduce QueryBandits, a bandit framework that designs rewrite strategies to maximize a reward model, that encapsulates hallucination propensity based upon the sensitivities of 17 linguistic features of the input query-and therefore, proactively steer LLMs away from generating hallucinations. Across 13 diverse QA benchmarks and 1,050 lexically perturbed queries per dataset, our top contextual QueryBandit (Thompson Sampling) achieves an 87.5% win rate over a no-rewrite baseline and also outperforms zero-shot static prompting ("paraphrase" or "expand") by 42.6% and 60.3% respectively. Therefore, we empirically substantiate the effectiveness of QueryBandits in mitigating hallucination via the intervention that takes the form of a query rewrite. Interestingly, certain static prompting strategies, which constitute a considerable number of current query rewriting literature, have a higher cumulative regret than the no-rewrite baseline, signifying that static rewrites can worsen hallucination. Moreover, we discover that the converged per-arm regression feature weight vectors substantiate that there is no single rewrite strategy optimal for all queries. In this context, guided rewriting via exploiting semantic features with QueryBandits can induce significant shifts in output behavior through forward-pass mechanisms, bypassing the need for retraining or gradient-based adaptation.

Related papers

• ADORE: Iterative Query Expansion with Retrieval-Grounded Relevance Feedback [22.297275934519345]
  We argue that effective expansion requires retrieval-grounded feedback, not just single-pass generation or unverified iteration.<n>We introduce ADORE, an iterative framework that turns retrieval outcomes into feedback for the next expansion.<n>Across TREC Deep Learning, BEIR, and BRIGHT, ADORE consistently outperforms strong query expansion baselines.
  arXiv  Detail & Related papers  (2026-06-11T20:52:37Z)
• Not All Queries Need Rewriting: When Prompt-Only LLM Refinement Helps and Hurts Dense Retrieval [0.11280931253550518]
  We show that prompt-only rewriting can be harmful in well-optimized verticals.<n>We identify a consistent mechanism: rewritten terms co-occur with reduced lexical alignment.<n>We suggest that domain-adaptive post-training is a safer strategy when supervision or implicit feedback is available.
  arXiv  Detail & Related papers  (2026-03-02T10:40:10Z)
• No One Size Fits All: QueryBandits for Hallucination Mitigation [27.117502300924183]
  We introduce QueryBandits, a model-agnostic contextual bandit framework that adaptively learns online to select the optimal query-rewrite strategy.<n>We show that there is no single rewrite policy optimal for all queries.<n>We also discover that certain static policies incur higher cumulative regret than No-Rewrite, indicating that an inflexible query-rewriting policy can worsen hallucinations.
  arXiv  Detail & Related papers  (2026-02-23T20:28:48Z)
• Reasoning-Augmented Representations for Multimodal Retrieval [27.4146940988752]
  Universal Multimodal Retrieval (UMR) seeks any-to-any search across text and vision.<n>We argue this brittleness is often data-specified: when images carry "silent" evidence and queries leave key semantics implicit, a single embedding pass must both reason and compress.<n>We propose a data-centric framework that decouples these roles by externalizing reasoning before retrieval.
  arXiv  Detail & Related papers  (2026-02-06T19:01:54Z)
• AFTER: Mitigating the Object Hallucination of LVLM via Adaptive Factual-Guided Activation Editing [32.581474255649894]
  Large Vision-Language Models (LVLMs) are susceptible to object hallucination.<n> editing the internal activations of LVLMs has shown promising effectiveness in mitigating hallucinations with minimal cost.<n>We propose Adaptive Factual-guided Visual-Textual Editing for hallucination mitigation.
  arXiv  Detail & Related papers  (2026-01-05T10:02:22Z)
• ReAG: Reasoning-Augmented Generation for Knowledge-based Visual Question Answering [54.72902502486611]
  ReAG is a Reasoning-Augmented Multimodal RAG approach that combines coarse- and fine-grained retrieval with a critic model that filters irrelevant passages.<n>ReAG significantly outperforms prior methods, improving answer accuracy and providing interpretable reasoning grounded in retrieved evidence.
  arXiv  Detail & Related papers  (2025-11-27T19:01:02Z)
• Chain-of-Code Collapse: Reasoning Failures in LLMs via Adversarial Prompting in Code Generation [0.3495246564946556]
  Large Language Models (LLMs) have achieved remarkable success in tasks requiring complex reasoning.<n>Do these models truly reason, or do they merely exploit shallow statistical patterns?<n>We introduce Chain-of-Code Collapse, where we investigate the robustness of reasoning LLMs by introducing a suite of semantically faithful yet adversarially structured prompt perturbations.
  arXiv  Detail & Related papers  (2025-06-08T02:43:46Z)
• Gumbel Reranking: Differentiable End-to-End Reranker Optimization [61.16471123356738]
  RAG systems rely on rerankers to identify relevant documents.<n> fine-tuning these models remains challenging due to the scarcity of annotated query-document pairs.<n>We propose Gumbel Reranking, an end-to-end training framework for rerankers aimed at minimizing the training-inference gap.
  arXiv  Detail & Related papers  (2025-02-16T13:23:39Z)
• R-Bot: An LLM-based Query Rewrite System [20.909806427953264]
  We propose R-Bot, an LLM-based query rewrite system with a systematic approach.<n>We first design a multi-source rewrite evidence preparation pipeline to generate query rewrite evidences.<n>We then propose a hybrid-semantics retrieval method that combines structural and semantic analysis.<n>We conduct comprehensive experiments on real-world datasets and widely used benchmarks, and demonstrate the superior performance of our system.
  arXiv  Detail & Related papers  (2024-12-02T16:13:04Z)
• Adaptive Query Rewriting: Aligning Rewriters through Marginal Probability of Conversational Answers [66.55612528039894]
  AdaQR is a framework for training query rewriting models with limited rewrite annotations from seed datasets and completely no passage label. A novel approach is proposed to assess retriever's preference for these candidates by the probability of answers conditioned on the conversational query.
  arXiv  Detail & Related papers  (2024-06-16T16:09:05Z)
• RaFe: Ranking Feedback Improves Query Rewriting for RAG [83.24385658573198]
  We propose a framework for training query rewriting models free of annotations. By leveraging a publicly available reranker, oursprovides feedback aligned well with the rewriting objectives.
  arXiv  Detail & Related papers  (2024-05-23T11:00:19Z)
• HyPoradise: An Open Baseline for Generative Speech Recognition with Large Language Models [81.56455625624041]
  We introduce the first open-source benchmark to utilize external large language models (LLMs) for ASR error correction. The proposed benchmark contains a novel dataset, HyPoradise (HP), encompassing more than 334,000 pairs of N-best hypotheses. LLMs with reasonable prompt and its generative capability can even correct those tokens that are missing in N-best list.
  arXiv  Detail & Related papers  (2023-09-27T14:44:10Z)
• Mutual Information Alleviates Hallucinations in Abstractive Summarization [73.48162198041884]
  We find a simple criterion under which models are significantly more likely to assign more probability to hallucinated content during generation: high model uncertainty. This finding offers a potential explanation for hallucinations: models default to favoring text with high marginal probability, when uncertain about a continuation. We propose a decoding strategy that switches to optimizing for pointwise mutual information of the source and target token--rather than purely the probability of the target token--when the model exhibits uncertainty.
  arXiv  Detail & Related papers  (2022-10-24T13:30:54Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.