mlsanityData to 7 Oct 2026

Vision Language Models Know Law of Conservation without Understanding More-or-Less

arXiv 2410.00332 ↗First posted 1 Oct 2024cs.AI q-bio.NCcited by 4

On the map: Multimodal LLMs / vision-language models → VLM benchmarks and visual reasoning

Abstract (arXiv)

Understanding law of conservation is a critical milestone in human cognitive development considered to be supported by the apprehension of quantitative concepts and the reversibility of operations. To assess whether this critical component of human intelligence has emerged in Vision Language Models, we have curated the ConserveBench, a battery of 365 cognitive experiments across four dimensions of physical quantities: volume, solid quantity, length, and number. The former two involve transformational tasks which require reversibility understanding. The latter two involve non-transformational tasks which assess quantity understanding. Surprisingly, we find that while Vision Language Models are generally good at transformational tasks, they tend to fail at non-transformational tasks. There is a dissociation between understanding the reversibility of operations and understanding the concept of quantity, which both are believed to be the cornerstones of understanding law of conservation in humans.

Key facts. Cited by 4 papers in the map (data through 7 Oct 2026). Placed in VLM benchmarks and visual reasoning, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 30 of its 62 references that are in the map. First posted on arXiv 1 Oct 2024.

Builds on

30 of 62 references.

Closest ancestors · 4

Same direction, newest first

  1. 2024CogVLM2: Visual Language Models for Image and Video Understanding29 Aug
  2. 2024Vision language models are blind: Failing to translate detailed visual features into words9 Jul
  3. 2023Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond24 Aug
  4. 2022Flamingo: a Visual Language Model for Few-Shot Learning29 Apr

Other references · 26

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2020Language Models are Few-Shot Learnerscited by 32,199
  3. 2023GPT-4 Technical Reportcited by 18,132
  4. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 13,958
  5. 2023Visual Instruction Tuningcited by 7,792
  6. 2023BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Modelscited by 6,069
  7. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,427
  8. 2023Improved Baselines with Visual Instruction Tuningcited by 4,362
  9. 2024Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolutioncited by 4,151
  10. 2023Qwen Technical Reportcited by 3,273
  11. 2016DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFscited by 3,239
  12. 2023MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Modelscited by 1,624
14 more
  1. 2019LXMERT: Learning Cross-Modality Encoder Representations from Transformerscited by 1,548
  2. 2019VisualBERT: A Simple and Performant Baseline for Vision and Languagecited by 1,291
  3. 2024SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilitiescited by 679
  4. 2023Instruction Tuning with GPT-4cited by 676
  5. 2023V*: Guided Visual Search as a Core Mechanism in Multimodal LLMscited by 459
  6. 2018Human Action Recognition and Prediction: A Surveycited by 226
  7. 2024SpatialBot: Precise Spatial Understanding with Vision Language Modelscited by 146
  8. 2022BridgeTower: Building Bridges Between Encoders in Vision-Language Representation Learningcited by 56
  9. 2016Going Deeper into Action Recognition: A Surveycited by 55
  10. 2022Consistent-Teacher: Towards Reducing Inconsistent Pseudo-targets in Semi-supervised Object Detectioncited by 43
  11. 2021Why AI is Harder Than We Thinkcited by 40
  12. 2021More than Encoder: Introducing Transformer Decoder to Upsamplecited by 26
  13. 2023Bridging the Domain Gap: Self-Supervised 3D Scene Understanding with Foundation Modelscited by 26
  14. 2023Diverse Cotraining Makes Strong Semi-Supervised Segmentorcited by 26

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 3; 2026: 1.

4 papers.

  1. 2025The Philosophical Foundations of Growing AI Like A Childcited by 6
  2. 2025Video Models Start to Solve Chess, Maze, Sudoku, Mental Rotation, and Raven' Matricescited by 4
  3. 2025Robusto-1 Dataset: Comparing Humans and VLMs on real out-of-distribution Autonomous Driving VQA from Perucited by 0
  4. 2026MindTopo: Can Foundation Models Reason in Topological Space?cited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Vision Language Models Cannot Reason About Physical Transformation
  2. 2025Decoupling the components of geometric understanding in Vision Language Models
  3. 2024Analyzing the Roles of Language and Vision in Learning from Limited Data
  4. 2025Reasoning Riddles: How Explainability Reveals Cognitive Limits in Vision-Language Models
  5. 2023Visual cognition in multimodal large language models
  6. 2026Unveiling the Visual Counting Bottleneck in Vision-Language Models
  7. 2025LLM world models are mental: Output layer evidence of brittle world model use in LLM mechanical reasoning
  8. 2026Knowledge Knows, Verbalization Tells: Disentangling Latent Directions for Mathematical Solvability in LLMs