mlsanityData to 7 Oct 2026

Conflict Adaptation in Vision-Language Models

arXiv 2510.24804 ↗First posted 28 Oct 2025cs.CV cs.CLcited by 1

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

Abstract (arXiv)

A signature of human cognitive control is conflict adaptation: improved performance on a high-conflict trial following another high-conflict trial. This phenomenon offers an account for how cognitive control, a scarce resource, is recruited. Using a sequential Stroop task, we find that 12 of 13 vision-language models (VLMs) tested exhibit behavior consistent with conflict adaptation, with the lone exception likely reflecting a ceiling effect. To understand the representational basis of this behavior, we use sparse autoencoders (SAEs) to identify task-relevant supernodes in InternVL 3.5 4B. Partially overlapping supernodes emerge for text and color in both early and late layers, and their relative sizes mirror the automaticity asymmetry between reading and color naming in humans. We further isolate a conflict-modulated supernode in layers 24-25 whose ablation significantly increases Stroop errors while minimally affecting congruent trials.

Key facts. Cited by 1 paper 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 13 of its 18 references that are in the map. First posted on arXiv 28 Oct 2025.

Builds on

13 of 18 references.

Closest ancestors · 3

Same direction, newest first

  1. 2025When Seeing Overrides Knowing: Disentangling Knowledge Conflicts in Vision-Language Models18 Jul
  2. 2025How Do Vision-Language Models Process Conflicting Information Across Modalities?2 Jul
  3. 2024Molmo and PixMo: Open Weights and Open Data for State-of-the-Art Vision-Language Models25 Sep

Other references · 10

  1. 2025Qwen3 Technical Reportcited by 8,195
  2. 2025Qwen2.5-VL Technical Reportcited by 5,418
  3. 2025Gemma 3 Technical Reportcited by 1,686
  4. 2025InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiencycited by 1,419
  5. 2023Sparse Autoencoders Find Highly Interpretable Features in Language Modelscited by 1,359
  6. 2024Transcoders Find Interpretable LLM Feature Circuitscited by 186
  7. 2024Auxiliary task demands mask the capabilities of smaller language modelscited by 54
  8. 2025Sparse Feature Coactivation Reveals Causal Semantic Modules in Large Language Modelscited by 6
  9. 2025Increasing Computation Resolves Conflicts in Vision Language Modelscited by 2
  10. 2024Do Language Models Understand the Cognitive Tasks Given to Them? Investigations with the N-Back Paradigmcited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Conflict and Congruency Effects in Large Language Models: In-Weight and In-Context Competition in a Verbal Conflict Taskcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Challenges in Understanding Modality Conflict in Vision-Language Models
  2. 2026Vision-Default, Prior-Override: Causal Mechanisms of Perception-Knowledge Conflict in Vision-Language Models
  3. 2026Arbitration Failure, Not Perceptual Blindness: How Vision-Language Models Resolve Visual-Linguistic Conflicts
  4. 2025Mixed Signals: Decoding VLMs' Reasoning and Underlying Bias in Vision-Language Conflict
  5. 2025Human Cognitive Benchmarks Reveal Foundational Visual Gaps in MLLMs
  6. 2026Revisiting Parameter Redundancy in Vision-Language-Action Models: Insights from VLM-to-VLA Adaptation
  7. 2026Which Source Wins? Task-Dependent Reliance in Vision-Language Models
  8. 2026Not All Features Are Created Equal: A Mechanistic Study of Vision-Language-Action Models