# SToRM: Supervised Token Reduction for Multi-modal LLMs toward efficient end-to-end autonomous driving

arXiv: https://arxiv.org/abs/2602.11656 · HTML: https://mlsanity.com/paper/2602.11656/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Driving VLMs and open-vocabulary detection, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 23 of its 24 references that are in the map.
- First posted on arXiv 12 Feb 2026.

## Placed in

- Multimodal LLMs / vision-language models / [Driving VLMs and open-vocabulary detection](https://mlsanity.com/mllm/driving-vlms-and-open-vocabulary-detection/)
- End-to-end autonomous driving / [Large language models for driving](https://mlsanity.com/driving/large-language-models-for-driving/)
- Efficient LLM inference and serving / [Visual token pruning in multimodal models](https://mlsanity.com/inference/visual-token-pruning-in-multimodal-models/)

## Builds on (most cited first, 23 of 23)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://mlsanity.com/paper/2301.12597/) (2023)
- [Neural Discrete Representation Learning](https://mlsanity.com/paper/1711.00937/) (2017)
- [Gaussian Error Linear Units (GELUs)](https://mlsanity.com/paper/1606.08415/) (2016)
- [Categorical Reparameterization with Gumbel-Softmax](https://mlsanity.com/paper/1611.01144/) (2016)
- [CARLA: An Open Urban Driving Simulator](https://mlsanity.com/paper/1711.03938/) (2017)
- [SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing](https://mlsanity.com/paper/1808.06226/) (2018)
- [MLP-Mixer: An all-MLP Architecture for Vision](https://mlsanity.com/paper/2105.01601/) (2021)
- [Token Merging: Your ViT But Faster](https://mlsanity.com/paper/2210.09461/) (2022)
- [DriveLM: Driving with Graph Visual Question Answering](https://mlsanity.com/paper/2312.14150/) (2023)
- [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://mlsanity.com/paper/2202.11094/) (2022)
- [LLaVA-PruMerge: Adaptive Token Reduction for Efficient Large Multimodal Models](https://mlsanity.com/paper/2403.15388/) (2024)
- [VisionZip: Longer is Better but Not Necessary in Vision Language Models](https://mlsanity.com/paper/2412.04467/) (2024)
- [LMDrive: Closed-Loop End-to-End Driving with Large Language Models](https://mlsanity.com/paper/2312.07488/) (2023)
- [DivPrune: Diversity-based Visual Token Pruning for Large Multimodal Models](https://mlsanity.com/paper/2503.02175/) (2025)
- [TinyLLaVA: A Framework of Small-scale Large Multimodal Models](https://mlsanity.com/paper/2402.14289/) (2024)
- [Vision Language Models in Autonomous Driving: A Survey and Outlook](https://mlsanity.com/paper/2310.14414/) (2023)
- [Recent Advancements in End-to-End Autonomous Driving using Deep Learning: A Survey](https://mlsanity.com/paper/2307.04370/) (2023)
- [ReasonNet: End-to-End Driving with Temporal and Global Reasoning](https://mlsanity.com/paper/2305.10507/) (2023)
- [Hybrid-Level Instruction Injection for Video Token Compression in Multi-modal Large Language Models](https://mlsanity.com/paper/2503.16036/) (2025)
- [Generalizing End-To-End Autonomous Driving In Real-World Environments Using Zero-Shot LLMs](https://mlsanity.com/paper/2411.14256/) (2024)

## Built on by (most cited first, 3 of 3)

- [LAD-Drive: Bridging Language and Trajectory with Action-Aware Diffusion Transformers](https://mlsanity.com/paper/2603.02035/) (2026)
- [Multi2: Hierarchical Multi-Agent Decision-Making with LLM-Based Agents in Interactive Environments](https://mlsanity.com/paper/2606.03698/) (2026)
- [Post-Hoc Merging is Not Enough: Many-Shot Model Merging with Loss-Gap Balancing](https://mlsanity.com/paper/2606.16501/) (2026)

## Abstract (arXiv)

In autonomous driving, end-to-end (E2E) driving systems that predict control commands directly from sensor data have achieved significant advancements. For safe driving in unexpected scenarios, these systems may additionally rely on human interventions such as natural language instructions. Using a multi-modal large language model (MLLM) facilitates human-vehicle interaction and can improve performance in such scenarios. However, this approach requires substantial computational resources due to its reliance on an LLM and numerous visual tokens from sensor inputs, which are limited in autonomous vehicles. Many MLLM studies have explored reducing visual tokens, but often suffer end-task performance degradation compared to using all tokens. To enable efficient E2E driving while maintaining performance comparable to using all tokens, this paper proposes the first Supervised Token Reduction framework for multi-modal LLMs (SToRM). The proposed framework consists of three key elements. First, a lightweight importance predictor with short-term sliding windows estimates token importance scores. Second, a supervised training approach uses an auxiliary path to obtain pseudo-supervision signals from an all-token LLM pass. Third, an anchor-context merging module partitions tokens into anchors and context tokens, and merges context tokens into relevant anchors to reduce redundancy while minimizing information loss. Experiments on the LangAuto benchmark show that SToRM outperforms state-of-the-art E2E driving MLLMs under the same reduced-token budget, maintaining all-token performance while reducing computational cost by up to 30x, and enabling real-time E2E driving on a standard GPU.