BALTO: Balanced Token-Level Policy Optimization for Hallucination Mitigation

• URL: http://arxiv.org/abs/2606.15893v2
• Date: Mon, 22 Jun 2026 03:05:35 GMT
• Title: BALTO: Balanced Token-Level Policy Optimization for Hallucination Mitigation
• Abstract summary: We propose a Balanced Token-level Policy Optimization framework for hallucination mitigation.<n>BalTO extracts checkable factual claims, verifies them against the reference context, and projects claim-level judgments to token-level labels.<n>Experiments on ConFiQA, RAGTruth, and FinLLM-Eval show that BALTO achieves the highest faithfulness across all six model--benchmark settings.
• Score: 51.22170603236523
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Hallucinations remain a major obstacle to deploying large language models (LLMs) in knowledge-intensive settings, where generated responses must be faithfully grounded in provided evidence. Reinforcement learning (RL) is a promising direction for hallucination mitigation, but response-level faithfulness rewards suffer from a granularity mismatch: localized hallucinations can cause supported content to receive spurious penalties. Although recent work introduces fine-grained feedback such as claim-level verification and token-level rewards, unbalanced credit assignment can still induce length, verbosity, or optimization-noise biases. We propose BALTO, a Balanced Token-level Policy Optimization framework for hallucination mitigation. BALTO extracts checkable factual claims, verifies them against the reference context, and projects claim-level judgments to token-level labels. A balanced token-level credit assignment mechanism is introduced into the framework. This design redistributes probability mass from unsupported content toward faithful content, rather than suppressing the entire response. We systematically analyze the limitations of response-level rewards from a theoretical standpoint, and prove BALTO's advantages in training stability and optimization efficiency for hallucination mitigation. Experiments on ConFiQA, RAGTruth, and FinLLM-Eval show that BALTO achieves the highest faithfulness across all six model--benchmark settings and consistently outperforms existing post-training baselines in Q-Score, demonstrating a stronger faithfulness--informativeness trade-off.

Related papers

• TRACE: A Unified Rollout Budget Allocation Framework for Efficient Agentic Reinforcement Learning [52.67756371231985]
  Reinforcement learning with verifiable rewards (RLVR) is a promising approach for enhancing reasoning and agentic behavior in large language models.<n>We introduce Tree Rollout Allocation for Contrastive Exploration (TRACE), a unified rollout allocation framework.<n>Technically, TRACE allocates rollout budget to both prompt roots and intermediate prefixes that are most likely to yield mixed terminal rewards.
  arXiv  Detail & Related papers  (2026-06-09T17:16:03Z)
• Towards Lightweight Reliability: Using Soft Prompts for Hallucination Mitigation in Large Language Models [5.019869660665509]
  Large language models (LLMs) have seen widespread adoption across various domains, yet their reliability is frequently undermined by hallucinations.<n>In high-stakes domains, these errors can reduce trust and introduce real-world risk.<n>We present a parameter-efficient approach that uses soft prompts to mitigate hallucinated content and promote responsible abstention.
  arXiv  Detail & Related papers  (2026-05-30T23:02:33Z)
• Verifiable Process Rewards for Agentic Reasoning [21.195739597726142]
  Reinforcement learning from verifiable rewards (RLVR) has improved the reasoning abilities of large language models (LLMs)<n>Most existing approaches rely on sparse outcome-level feedback.<n>We propose Verifiable Process Rewards (VPR), a framework that converts such oracles into dense turn-level supervision for reinforcement learning.
  arXiv  Detail & Related papers  (2026-05-11T10:30:53Z)
• I-CALM: Incentivizing Confidence-Aware Abstention for LLM Hallucination Mitigation [2.688126466709795]
  Large language models (LLMs) frequently produce confident but incorrect answers.<n>We study whether prompt-only interventions can reduce hallucination risk without modifying the model.<n>Results show the framework can improve selective answering on factual questions without retraining.
  arXiv  Detail & Related papers  (2026-04-05T00:15:41Z)
• Balancing Faithfulness and Performance in Reasoning via Multi-Listener Soft Execution [79.98699884805636]
  Reasoning Execution by Multiple Listeners (REMUL) is a multi-party reinforcement learning approach.<n>REMUL builds on the hypothesis that reasoning traces which other parties can follow will be more faithful.<n>Speakers are rewarded for producing reasoning that is clear to listeners.
  arXiv  Detail & Related papers  (2026-02-18T02:55:55Z)
• Towards Generalizable Reasoning: Group Causal Counterfactual Policy Optimization for LLM Reasoning [50.352417879912515]
  Large language models (LLMs) excel at complex tasks with advances in reasoning capabilities.<n>We propose Group Causal Counterfactual Policy Optimization to explicitly train LLMs to learn generalizable reasoning patterns.<n>We then construct token-level advantages from this reward and optimize the policy, encouraging LLMs to favor reasoning patterns that are process-valid and counterfactually robust.
  arXiv  Detail & Related papers  (2026-02-06T08:03:11Z)
• Stop Rewarding Hallucinated Steps: Faithfulness-Aware Step-Level Reinforcement Learning for Small Reasoning Models [59.6715047267181]
  Small reasoning models (SRMs) are prone to hallucinations, especially in intermediate reasoning steps.<n>Existing mitigation methods based on online reinforcement learning rely on outcome-based rewards or coarse-grained chain-of-thought evaluation.<n>We propose Faithfulness-Aware Step-Level Reinforcement Learning (FaithRL), introducing step-level supervision via explicit faithfulness rewards from a process reward model.
  arXiv  Detail & Related papers  (2026-02-05T17:15:12Z)
• Step Potential Advantage Estimation: Harnessing Intermediate Confidence and Correctness for Efficient Mathematical Reasoning [25.562101968892833]
  Reinforcement Learning with Verifiable Rewards (RLVR) elicits long chain-of-thought reasoning in large language models (LLMs)<n>Existing approaches improve RLVR via token-level entropy or sequence-level length control, but lack a semantically grounded, step-level measure of reasoning progress.<n>We propose Step Potential Advantage Estimation (SPAE), a fine-grained credit assignment method that amplifies potential gains, penalizes potential drops, and applies penalty after potential saturates to encourage timely termination.
  arXiv  Detail & Related papers  (2026-01-07T11:36:01Z)
• Information-Theoretic Reward Modeling for Stable RLHF: Detecting and Mitigating Reward Hacking [78.69179041551014]
  We propose an information-theoretic reward modeling framework based on the Information Bottleneck principle.<n>We show that InfoRM filters out preference-irrelevant information to alleviate reward misgeneralization.<n>We also introduce IBL, a distribution-level regularization that penalizes such deviations, effectively expanding the optimization landscape.
  arXiv  Detail & Related papers  (2025-10-15T15:51:59Z)
• Reasoning Models Hallucinate More: Factuality-Aware Reinforcement Learning for Large Reasoning Models [83.24079543652253]
  Large language models (LLMs) have significantly advanced in reasoning tasks through reinforcement learning (RL) optimization.<n>However, reasoning-oriented RL fine-tuning significantly increases the prevalence of hallucinations.<n>We propose Factuality-aware Step-wise Policy Optimization (FSPO), an innovative RL fine-tuning algorithm incorporating explicit factuality verification.
  arXiv  Detail & Related papers  (2025-05-30T14:23:32Z)

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.