Seeing Through the Chain: Understanding and Mitigating Hallucinations in Multimodal Large Reasoning Models

Hao Fang; Jinyu Li; Jiawei Kong; Tianqu Zhuang; Kuofeng Gao; Bin Chen; Shu-Tao Xia

cs.CVFeb 3, 2026

Seeing Through the Chain: Understanding and Mitigating Hallucinations in Multimodal Large Reasoning Models

Authors: Hao Fang, Jinyu Li, Jiawei Kong, Tianqu Zhuang, Kuofeng Gao, Bin Chen, Shu-Tao Xia

Organizations: Tsinghua Shenzhen International Graduate School, Tsinghua University · Harbin Institute of Technology, Shenzhen

Abstract

While multimodal reasoning models (MLRMs) have exhibited impressive capabilities, they remain prone to hallucinations, and effective solutions are still underexplored. In this paper, we experimentally analyze the hallucination cause and propose C3PO, a training-based mitigation framework comprising Chain-of-Thought Compression and Contrastive Preference Optimization. Firstly, we identify that introducing reasoning mechanisms exacerbates models' reliance on language priors while overlooking visual inputs, which can produce CoTs with reduced visual cues but redundant text tokens. To this end, we propose to selectively filter redundant thinking tokens for a more compact and signal-efficient CoT representation that preserves task-relevant information while suppressing noise. In addition, we observe that the quality of the reasoning trace largely determines whether hallucination emerges in subsequent responses. To leverage this insight, we introduce a reasoning-enhanced preference tuning scheme that constructs training pairs using high-quality AI feedback. We further design a multimodal hallucination-inducing mechanism that elicits models' inherent hallucination patterns via carefully crafted inducers, yielding informative negative signals for contrastive correction. We provide theoretical justification for the effectiveness and demonstrate consistent hallucination reduction across diverse MLRMs and benchmarks.

Figures & tables

Figure 1: (a) A hallucination case from R1-Onevision Yang et al. (2025a) . (b) Performance of MLRMs R1-Onevision and MM-Eureka Meng et al. (2025) , and the base model Qwen2.5-VL-7B Bai et al. (2025b) on AMBER Wang et al. (2023) , where high values indicate fewer hallucinations . The hallucination increases from the base model to reasoning variants.

Figure 2: Two important hallucination analyses for MLRMs. (a) Attention distributions of MLRMs and the non-reasoning Qwen2.5-VL on 300 samples from MSCOCO Lin et al. (2014) . As observed, MLRMs exhibit a substantial reduction in attention to visual tokens. (b) Proportion of hallucinated answers given hallucinated (-H) and non-hallucinated (-NH) CoTs under CHAIR Rohrbach et al. (2018) evaluation. EU denotes MM-Eureka, and OV denotes R1-Onevision. The result of Qwen2.5-VL is reported as a reference.

Figure 3: Overview of the proposed C3PO framework. We first construct SFT datasets by removing redundant low-importance tokens in generated CoTs. Next, we perform hallucination-aware preference optimization, where preferred samples contain CoTs enhanced by an advanced open-source MLLM, while the original unmodified outputs serve as rejected ones. To further enhance the optimization, we propose a multimodal hallucination-inducing technique that crafts degraded visual inputs and misleading prompts to elicit MLRM’s inherent hallucination patterns as negative contrasts.

Method	R1-Onevision		Orsta-R1		MM-Eureka		MM-R1		ThinkLite
Method	$C_{S}$ $\downarrow$	$C_{I}$ $\downarrow$	$C_{S}$ $\downarrow$	$C_{I}$ $\downarrow$	$C_{S}$ $\downarrow$	$C_{I}$ $\downarrow$	$C_{S}$ $\downarrow$	$C_{I}$ $\downarrow$	$C_{S}$ $\downarrow$	$C_{I}$ $\downarrow$
Vanilla	36.7	8.2	40.7	8.5	44.0	8.6	37.0	7.4	41.7	7.8
RLAIF-V	39.0	8.3	39.7	7.6	41.3	8.5	38.7	7.5	40.3	7.5
OPA-DPO	36.7	7.7	41.7	8.2	41.3	7.7	36.3	7.2	40.0	7.9
C3PO	35.0	7.4	35.3	6.9	40.0	7.7	35.0	7.2	38.0	7.3

Table 1: Comparison of the proposed C3PO with competitive baselines on the CHAIR benchmark. We evaluate on images from MSCOCO.

C_{S}

and

C_{I}

denotes CHAIR S and CHAIR I respectively.

Model	Method	Random		Popular		Adversarial
Model	Method	Acc.	F1 score	Acc.	F1 score	Acc.	F1 score
MM-Eureka	Vanilla	84.12	81.90	83.90	81.23	83.47	80.85
	RLAIF-V	83.81	81.46	83.70	80.89	83.27	80.47
	OPA-DPO	85.29	83.46	84.93	82.70	84.37	82.16
	C3PO	86.22	84.68	85.87	83.94	85.23	83.34
MM-R1	Vanilla	84.85	82.83	84.47	82.04	83.97	81.56

Table 2: Comparison of the proposed C3PO with competitive baselines on the POPE benchmark (%). We report the accuracy and F1 score for evaluation.

Figure 4: GPT-4 assisted benchmark. Sentence-level Hallucination Ratio (SHR) measures the hallucination degree. We also provide 1&2-gram, the number of sentences per image (SPI), and words per image (WPI). A larger radar area indicates better performance.

Method	R1-Onevision		Orsta-R1		MM-Eureka		MM-R1		ThinkLite
Method	Acc.	F1 score	Acc.	F1 score	Acc.	F1 score	Acc.	F1 score	Acc.	F1 score
Vanilla	81.52	86.54	82.94	87.31	83.27	87.15	84.38	88.20	82.25	86.96
RLAIF-V	80.99	84.97	83.14	87.50	83.46	87.49	84.17	88.11	82.31	87.04
OPA-DPO	80.71	85.30	83.58	87.67	83.72	87.59	84.76	88.51	83.02	87.36
C3PO	81.90	86.73	84.57	88.50	85.01	88.80	85.61	89.25	83.96	88.26

Table 3: Comparison of the proposed C3PO with competitive baselines on the AMBER benchmark (%). We report the accuracy and F1 score for evaluation.

Method	CHAIR Eval		GPT-4 Eval
Method	$C_{S}$ $\downarrow$	$C_{I}$ $\downarrow$	SHR $\downarrow$
C3PO	35.3	6.9	26.3
w/o loss $\mathcal{L}_{\mathrm{RE}}$	37.3	7.2	28.8
w/o loss $\mathcal{L}_{\mathrm{DPO}}$	39.7	8.1	29.4
w/o loss $\mathcal{L}_{\mathrm{Anc}}$	37.7	7.5	27.2
w/o MHI strategy	37.0	7.9	28.4

Table 4: Ablation study of the proposed techniques in C3PO under CHAIR and GPT-4 assisted evaluation benchmarks.

Appendix figures & tables7 assets

Supplementary material from the paper’s appendix.

Appendix

Model	Method	Random		Popular		Adversarial
Model	Method	Accuracy	F1 score	Accuracy	F1 score	Accuracy	F1 score
R1-Onevision	Vanilla	83.64	81.82	82.80	80.59	81.93	79.81
	RLAIF-V	82.85	81.03	81.60	79.45	81.57	79.42
	OPA-DPO	83.85	82.28	82.57	80.67	82.03	80.19
	C3PO	83.88	82.66	82.53	81.01	82.37	80.87
Orsta-R1	Vanilla	85.15	83.29	84.70	82.43	84.13	81.90

Appendix

Table 5: Additional performance comparison of the proposed C3PO and strong baselines on the POPE benchmark. We supplement results on MLRMs Orsta-R1 and R1-Onevision.

Figure 7: GPT-4 assisted benchmark results for R1-Onevision. A larger radar area indicates better general performance.

Metric	R1-Onevision			Orsta-R1			MM-Eureka			MM-R1			ThinkLite
Metric	C3PO	w/o DPO	Vanilla	C3PO	w/o DPO	Vanilla	C3PO	w/o DPO	Vanilla	C3PO	w/o DPO	Vanilla	C3PO	w/o DPO	Vanilla
$C_{S}$ $\downarrow$	35.0	40.0	36.7	35.3	39.7	40.7	40.0	40.7	44.0	35.0	35.3	37.0	38.0	41.3	41.7
$C_{I}$ $\downarrow$	7.4	7.7	8.2	6.9	7.7	8.5	7.7	8.3	8.6	7.2	6.9	7.4	7.3	7.2	7.8
SHR $\downarrow$	37.9	36.7	40.7	26.3	29.2	30.6	28.1	29.3	30.6	27.9	31.0	29.8	28.5	29.9	32.1

Appendix

Table 6: Ablation study of C3PO on the preference optimization and the CoT compression.

Figure 9: Visualization of the pruned CoTs.

Figure 10: Visualization results comparing our C3PO and other methods. Hallucinations are marked in red .

Figure 11: Visualization results comparing our C3PO and other methods. Hallucinations are marked in red .

Hallucination Mitigation

Hallucination mitigation develops methods to detect and reduce unsupported AI outputs while preserving useful, evidence-grounded responses.

See all 27 papers

Reasoning Matters: Mitigate Hallucination in Multimodal Large Reasoning Models via Reasoning-Conditioned Preference Optimization
May 27, 2026Jiawei Kong, Hao Fang, Shunxiang Liao +5Multimodal Grounding Direct Preference Optimization
Tsinghua Shenzhen International Graduate School, Tsinghua University · Harbin Institute of Technology, Shenzhen
Understanding and Mitigating Hallucinations in Multimodal Chain-of-Thought Models
Mar 28, 2026Ji Ma, Wei Suo, Peng Wang +1Multimodal Hallucination Multimodal CoT Reasoning
School of Computer Science and Ningbo Institute, Northwestern Polytechnical University, China. · National Engineering Laboratory for Integrated Aero-Space-Ground-Ocean Big Data Application Technology, China.
Cognitive Pivot Points and Visual Anchoring: Unveiling and Rectifying Hallucinations in Multimodal Reasoning Models
Apr 11, 2026Zhe Qian, Yanbiao Ma, Zhuohan Ouyang +7VLM Hallucination Mitigation Multimodal Reasoning
South China Agricultural University · Gaoling School of Artificial Intelligence, Renmin University of China · South China Normal University +4

May 27, 2026J/K move · Enter open · S save

Reasoning Matters: Mitigate Hallucination in Multimodal Large Reasoning Models via Reasoning-Conditioned Preference Optimization

Jiawei Kong, Hao Fang, Shunxiang Liao, Jinyu Li+4

Tsinghua Shenzhen International Graduate School, Tsinghua University · Harbin Institute of Technology, Shenzhen

Multimodal Grounding Direct Preference Optimization Hallucination in Language Models Multimodal Large Language Models Multimodal Hallucination Multimodal CoT Reasoning

Open paper AskPDF arXiv

Multimodal Large Reasoning Models introduce the reasoning paradigm, demonstrating strong capabilities on complex vision-language tasks. However, they still suffer from severe hallucinations. Existing training-based methods typically mitigate hallucinations through response-level direct preference optimization (DPO), where the Chain-of-Thought (CoT) and the final answer are treated as a monolithic output and optimized jointly. We reveal that this formulation performs similarly to answer-only optimization, suggesting that it primarily learns answer-level preference, while leaving CoT-level supervision insufficiently exploited. To address this issue, we explicitly formulate a CoT-oriented preference term and derive Reasoning-Conditioned Direct Preference Optimization (RC-DPO), which models the CoT as a condition for answer generation and contrasts the preference for the same preferred answer under different CoT conditions, promoting answer-supportive reasoning chain alignment. To further improve optimization, we introduce a reasoning-enhanced preference data generation strategy that employs Monte Carlo Tree Search to discover visually grounded and logically consistent CoTs as positive samples, and attention-guided CoT token pruning to construct negative ones. Extensive experiments across various models and benchmarks show that RC-DPO effectively mitigates hallucinations and improves the reliability of the multimodal reasoning process.

Abstract

Hallucination Mitigation

Explore similar work

Reasoning Matters: Mitigate Hallucination in Multimodal Large Reasoning Models via Reasoning-Conditioned Preference Optimization

Understanding and Mitigating Hallucinations in Multimodal Chain-of-Thought Models

Cognitive Pivot Points and Visual Anchoring: Unveiling and Rectifying Hallucinations in Multimodal Reasoning Models