Trace, Verify, and Correct: A Training-Free Framework for Spatial Reasoning in Multimodal LLMs
Yang Yang, Jiawei Chen, Tairan Chen, Zhaoxia Yin
cs.CV, cs.CL
Submitted: 2026-08-19
Updated: 2026-08-20
Comments: 19 pages, 7 figures
License: http://creativecommons.org/licenses/by/4.0/
The gist: Although Multimodal Large Language Models (MLLMs) have made substantial progress, their spatial reasoning may still produce intermediate judgments inconsistent with the input image, allowing errors
Terminology
Abstract
Although Multimodal Large Language Models (MLLMs) have made substantial progress, their spatial reasoning may still produce intermediate judgments inconsistent with the input image, allowing errors to propagate through the reasoning chain and affect the final answer. Existing methods mainly improve spatial reasoning through training or additional spatial information, without considering whether the reasoning process itself is faithful to the model input. Our study shows that unfaithful reasoning chains significantly reduce final-answer accuracy. To address this issue, we propose a modular and training-free framework for spatial reasoning verification and correction. The framework constructs a Spatial Evidence Graph (SEG), which associates atomic spatial evidence extracted from Chain-of-Thought reasoning with visual entities, spatial relations, source steps, and visual evidence. Spatial Evidence Reliability Assessment (SERA) evaluates the reliability of visual evidence based on object existence, localization, and geometric measurements. The framework then identifies the earliest spatial evidence unit contradicted by reliable visual evidence and guides the original MLLM to revise the subsequent reasoning and final answer. Across 15 model-dataset settings, our method achieves an average accuracy of 68.94%, outperforming the compared baselines by 8.55 percentage points on average. Our code will be open-sourced.
Sources
- Invisible to Humans, Triggered by Agents: Stealthy Jailbreak Attacks on Mobile Vision-Language Agents
- Graph-of-Mark: Promote Spatial Reasoning in Multimodal Language Models with Graph-Based Visual Prompting
- Spatial Reasoning with Vision-Language Models in Ego-Centric Multi-View Scenes
- The Llama 3 Herd of Models
- Qwen3-VL Technical Report
- Tex3D: Objects as Attack Surfaces via Adversarial 3D Textures for Vision-Language-Action Models
- M3CoTBench: Benchmark Chain-of-Thought of MLLMs in Medical Image Understanding
- SPD-Faith Bench: Diagnosing and Improving Faithfulness in Chain-of-Thought for Multimodal Large Language Models
- REVISOR: Beyond Textual Reflection, Towards Multimodal Introspective Reasoning in Long-Form Video Understanding
- Beyond Unimodal Shortcuts: MLLMs as Cross-Modal Reasoners for Grounded Named Entity Recognition
- See&Trek: Training-Free Spatial Prompting for Multimodal Large Language Model
- MLLM can see? Dynamic Correction Decoding for Hallucination Mitigation
- InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
- N3D-VLM: Native 3D Grounding Enables Accurate Spatial Reasoning in Vision-Language Models
- Set-of-Mark Prompting Unleashes Extraordinary Visual Grounding in GPT-4V
- CoRGI: Verified Chain-of-Thought Reasoning with Post-hoc Visual Grounding
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models