# Do LiDAR Language Models Really Understand Spatio-temporal Relationships?

arXiv: https://arxiv.org/abs/2609.24452 · HTML: https://mlsanity.com/paper/2609.24452/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 26 of its 29 references that are in the map.
- First posted on arXiv 21 Sep 2026.

## Builds on (most cited first, 26 of 26)

- [nuScenes: A multimodal dataset for autonomous driving](https://mlsanity.com/paper/1903.11027/) (2019)
- [Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answering](https://mlsanity.com/paper/1612.00837/) (2016)
- [Evaluating Object Hallucination in Large Vision-Language Models](https://mlsanity.com/paper/2305.10355/) (2023)
- [Mitigating Object Hallucinations in Large Vision-Language Models through Visual Contrastive Decoding](https://mlsanity.com/paper/2311.16922/) (2023)
- [Object Hallucination in Image Captioning](https://mlsanity.com/paper/1809.02156/) (2018)
- [When and why vision-language models behave like bags-of-words, and what to do about it?](https://mlsanity.com/paper/2210.01936/) (2022)
- [DriveLM: Driving with Graph Visual Question Answering](https://mlsanity.com/paper/2312.14150/) (2023)
- [Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality](https://mlsanity.com/paper/2204.03162/) (2022)
- [DriveGPT4: Interpretable End-to-end Autonomous Driving via Large Language Model](https://mlsanity.com/paper/2310.01412/) (2023)
- [Don't Just Assume; Look and Answer: Overcoming Priors for Visual Question Answering](https://mlsanity.com/paper/1712.00377/) (2017)
- [AMBER: An LLM-free Multi-dimensional Benchmark for MLLMs Hallucination Evaluation](https://mlsanity.com/paper/2311.07397/) (2023)
- [Large Language Models Sensitivity to The Order of Options in Multiple-Choice Questions](https://mlsanity.com/paper/2308.11483/) (2023)
- [NuScenes-QA: A Multi-modal Visual Question Answering Benchmark for Autonomous Driving Scenario](https://mlsanity.com/paper/2305.14836/) (2023)
- [LingoQA: Visual Question Answering for Autonomous Driving](https://mlsanity.com/paper/2312.14115/) (2023)
- [LiDAR-LLM: Exploring the Potential of Large Language Models for 3D LiDAR Understanding](https://mlsanity.com/paper/2312.14074/) (2023)
- [VidHalluc: Evaluating Temporal Hallucinations in Multimodal Large Language Models for Video Understanding](https://mlsanity.com/paper/2412.03735/) (2024)
- [EventHallusion: Diagnosing Event Hallucinations in Video LLMs](https://mlsanity.com/paper/2409.16597/) (2024)
- [NuScenes-SpatialQA: A Spatial Understanding and Reasoning Benchmark for Vision-Language Models in Autonomous Driving](https://mlsanity.com/paper/2504.03164/) (2025)
- [LidarCLIP or: How I Learned to Talk to Point Clouds](https://mlsanity.com/paper/2212.06858/) (2022)
- [Unveiling the Mist over 3D Vision-Language Understanding: Object-centric Evaluation with Chain-of-Analysis](https://mlsanity.com/paper/2503.22420/) (2025)
- [Strengthened Symbol Binding Makes Large Language Models Reliable Multiple-Choice Selectors](https://mlsanity.com/paper/2406.01026/) (2024)
- [Do 3D Large Language Models Really Understand 3D Spatial Relationships?](https://mlsanity.com/paper/2603.23523/) (2026)
- [4D-Bench: Benchmarking Multi-modal Large Language Models for 4D Object Understanding](https://mlsanity.com/paper/2503.17827/) (2025)
- [Spatial4D-Bench: A Versatile 4D Spatial Intelligence Benchmark](https://mlsanity.com/paper/2601.00092/) (2025)
- [B4DL: A Benchmark for 4D LiDAR LLM in Spatio-Temporal Understanding](https://mlsanity.com/paper/2508.05269/) (2025)
- [DRIVESPATIAL: A Benchmark for Spatiotemporal Intelligence in VLMs for Autonomous Driving](https://mlsanity.com/paper/2605.23176/) (2026)

## Abstract (arXiv)

Recent 4D LiDAR language models aim to reason about objects and their evolving spatial relationships. Yet, in our evaluation, always selecting the same option nearly matches the multiple-choice accuracy of two B4DL-derived configurations. We introduce LiDAR-Hallu, a geometry-referenced benchmark and diagnostic protocol with 10,000 questions across 150 nuScenes scenes. It covers object existence, ego-relative position, distance ordering, relative motion, and temporal localization, with explicit rules for selecting objects, comparing times, and determining reference answers. Our protocol combines fixed-answer and candidate-content controls, cross-scene pairs with identical prompts but opposite reference answers, and relation-specific recall. Analysis of 100,000 recorded responses reveals failures hidden by aggregate accuracy. Candidate duration alone makes temporal answers predictable without observing LiDAR. On paired questions, the models frequently give the same answer to scenes requiring opposite answers. Relation-specific analysis further shows that both configurations miss every positive lateral-motion case across all tested conditions. Temporal-shuffle contrastive decoding provides little net improvement, as repairs are largely offset by new errors and the main failures persist. These results show that evaluating spatio-temporal reasoning requires testing whether models distinguish the queried physical relationships, rather than relying on individual-answer accuracy alone. The source code, checkpoints, and data are released at https://github.com/Awesome4D/4DMLLM_Hallucination_Bench.