# Revealing Interpretable Failure Modes of VLMs

arXiv: https://arxiv.org/abs/2605.12674 · HTML: https://mlsanity.com/paper/2605.12674/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Chain-of-thought reasoning for driving, one of 17 directions in LLM reasoning.
- Builds on 22 of its 37 references that are in the map.
- First posted on arXiv 12 May 2026.

## Placed in

- LLM reasoning / [Chain-of-thought reasoning for driving](https://mlsanity.com/lr/chain-of-thought-reasoning-for-driving/)
- Multimodal LLMs / vision-language models / [Driving VLMs and open-vocabulary detection](https://mlsanity.com/mllm/driving-vlms-and-open-vocabulary-detection/)

## Builds on (most cited first, 22 of 22)

- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities](https://mlsanity.com/paper/2507.06261/) (2025)
- [CARLA: An Open Urban Driving Simulator](https://mlsanity.com/paper/1711.03938/) (2017)
- [Why Do Multi-Agent LLM Systems Fail?](https://mlsanity.com/paper/2503.13657/) (2025)
- [VLMEvalKit: An Open-Source Toolkit for Evaluating Large Multi-Modality Models](https://mlsanity.com/paper/2407.11691/) (2024)
- [A Comprehensive Survey of Scene Graphs: Generation and Application](https://mlsanity.com/paper/2104.01111/) (2021)
- [Are VLMs Ready for Autonomous Driving? An Empirical Study from the Reliability, Data, and Metric Perspectives](https://mlsanity.com/paper/2501.04003/) (2025)
- [Concept-based Explainable Artificial Intelligence: A Survey](https://mlsanity.com/paper/2312.12936/) (2023)
- [Scenic: A Language for Scenario Specification and Scene Generation](https://mlsanity.com/paper/1809.09310/) (2018)
- [Explain Any Concept: Segment Anything Meets Concept-Based Explanation](https://mlsanity.com/paper/2305.10289/) (2023)
- [ManipBench: Benchmarking Vision-Language Models for Low-Level Robot Manipulation](https://mlsanity.com/paper/2505.09698/) (2025)
- [ADVEDM:Fine-grained Adversarial Attack against VLM-based Embodied Agents](https://mlsanity.com/paper/2509.16645/) (2025)
- [Failure Modes in Machine Learning Systems](https://mlsanity.com/paper/1911.11034/) (2019)
- [On the Vulnerability of LLM/VLM-Controlled Robotics](https://mlsanity.com/paper/2402.10340/) (2024)
- [VLM-RobustBench: A Comprehensive Benchmark for Robustness of Vision-Language Models](https://mlsanity.com/paper/2603.06148/) (2026)
- [Failure Modes in LLM Systems: A System-Level Taxonomy for Reliable AI Applications](https://mlsanity.com/paper/2511.19933/) (2025)
- [Large Language Model-assisted Autonomous Vehicle Recovery from Immobilization](https://mlsanity.com/paper/2510.26023/) (2025)
- [Best Subset Selection: Optimal Pursuit for Feature Selection and Elimination](https://mlsanity.com/paper/2501.16815/) (2025)
- [Toward Inherently Robust VLMs Against Visual Perception Attacks](https://mlsanity.com/paper/2506.11472/) (2025)
- [ScenicRules: An Autonomous Driving Benchmark with Multi-Objective Specifications and Abstract Scenarios](https://mlsanity.com/paper/2602.16073/) (2026)
- [Discovering Failure Modes in Vision-Language Models using RL](https://mlsanity.com/paper/2604.04733/) (2026)

## Abstract (arXiv)

Vision-Language Models (VLMs) are increasingly used in safety-critical applications because of their broad reasoning capabilities and ability to generalize with minimal task-specific engineering. Despite these advantages, they can exhibit catastrophic failures in specific real-world situations, constituting failure modes. We introduce REVELIO, a framework for systematically uncovering interpretable failure modes in VLMs. We define a failure mode as a composition of interpretable, domain-relevant concepts-such as pedestrian proximity or adverse weather conditions-under which a target VLM consistently behaves incorrectly. Identifying such failures requires searching over an exponentially large discrete combinatorial space. To address this challenge, REVELIO combines two search procedures: a diversity-aware beam search that efficiently maps the failure landscape, and a Gaussian-process Thompson Sampling strategy that enables broader exploration of complex failure modes. We apply REVELIO to autonomous driving and indoor robotics domains, uncovering previously unreported vulnerabilities in state-of-the-art VLMs. In driving environments, the models often demonstrate weak spatial grounding and fail to account for major obstructions, leading to recommendations that would result in simulated crashes. In indoor robotics tasks, VLMs either miss safety hazards or behave excessively conservatively, producing false alarms and reducing operational efficiency. By identifying structured and interpretable failure modes, REVELIO offers actionable insights that can support targeted VLM safety improvements.