pith. sign in

arxiv: 2606.20177 · v2 · pith:4BOGZULDnew · submitted 2026-06-18 · 💻 cs.CV · cs.AI

Evaluating and Enhancing Negation Comprehension in Remote Sensing MLLMs

Haochen Han , Jue Wang , Alex Jinpeng Wang , Fangming Liu This is my paper

Pith reviewed 2026-06-26 18:06 UTC · model grok-4.3

classification 💻 cs.CV cs.AI
keywords remote sensingnegation comprehensionmultimodal large language modelsbenchmarktest-time learninghallucinationsimage understanding
0
0 comments X

The pith

Remote sensing MLLMs struggle with negation in imagery and exhibit hallucinations, yet NeFo recovers much of the lost performance with about 5% unlabeled test samples.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that advanced multimodal large language models for remote sensing cannot reliably process statements about what is absent or false in an image. It demonstrates this limitation through a new benchmark called RS-Neg that spans region-level to scene-level tasks and uses an automated pipeline to generate diverse negation queries. Evaluation on the benchmark shows clear performance drops and incorrect affirmations of negated content. The authors then introduce NeFo, a test-time learning approach that incorporates the logical role of negation during optimization. A sympathetic reader would care because real applications such as emergency response require accurate detection of missing or non-occurring features.

Core claim

Advanced RS MLLMs struggle with negation, exhibiting hallucinations and substantial performance degradation. To close this gap, NeFo is proposed as a novel test-time learning method that explicitly incorporates the logical role of negation into the model optimization. Using about 5% unlabeled test samples, NeFo significantly improves the negation understanding of models and shows strong generalization to unseen tasks.

What carries the argument

NeFo, a test-time learning method that incorporates the logical role of negation into model optimization to reduce hallucinations.

Load-bearing premise

The automated pipeline that uses LLMs to synthesize negation queries, together with the dynamic visual focus module, produces queries that represent real negation comprehension needs in remote sensing imagery.

What would settle it

A human evaluation in which domain experts rate the realism of the generated RS-Neg queries or measure whether NeFo produces no gains on a manually authored negation test set.

Figures

Figures reproduced from arXiv: 2606.20177 by Alex Jinpeng Wang, Fangming Liu, Haochen Han, Jue Wang.

Figure 1
Figure 1. Figure 1 [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Pipeline for constructing RS-Neg dataset, with MCQ as an example. Given an RS image-caption pair, we first use an LLM to extract concepts from the caption and generate corresponding negative counterparts. Second, we employ an MCTS-based visual focus method to verify these negative concepts against the image. Third, we use an LLM to formulate task-specific samples and enhance linguistic diversity. to verify… view at source ↗
Figure 3
Figure 3. Figure 3: Model performance on RS-Neg across different MLLMs and tasks. (a) and (b) show the performance drop compared to original queries on VQA and classification tasks, respectively. (c) reports the results on the visual grounding task, where we include the performance of Qwen2.5-VL on corresponding affirmative queries as a positive baseline. (d) reports the results on the MCQ task, where several RS-specific mode… view at source ↗
Figure 4
Figure 4. Figure 4: An overview of the proposed NeFo. Given the test sample with a negation query, we first use a lexicon to construct the negation-masked text counterpart. Then we use LoRA to fine-tune MLLMs with two training objectives: the Truth-Value Inversion Loss that forces the model to focus on negation-related information, and the Knowledge Retaining Loss that preserves the foundational knowledge on negation-masked v… view at source ↗
Figure 5
Figure 5. Figure 5: Sensitivity analysis of hyper-parameters in NeFo. (a) RS-Neg VQA with RS￾LLaVA as the base model. (b) RS-Neg MCQ with GeoReason as the base model [PITH_FULL_IMAGE:figures/full_fig_p012_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Effect of NeFo with varying test-time adaptation data sizes. We report IoU@0.5 for RS-Neg Grounding and accuracy for the remaining tasks. iments follow the same training settings for a fair comparison. We increase the training data for Qwen2.5-VL-7B on RS-Neg MCQ from 50 to 700 samples and evaluate its zero-shot performance on RS-Neg Classification, RS-Neg Grounding, and FloodNet VQA. As shown in [PITH_FU… view at source ↗
read the original abstract

Multimodal Large Language Models (MLLMs) have demonstrated remarkable success in various Remote Sensing (RS) tasks. However, their ability to comprehend negation remains underexplored, limiting deployment in real-world applications where models must explicitly identify what is false or absent, e.g., emergency responders need to locate non-flooded routes for evacuation. To comprehensively study this limitation, we introduce RS-Neg, the first benchmark to evaluate negation understanding across region-level to scene-level tasks. Specifically, we design an automated data generation pipeline for RS imagery, using LLMs to synthesize diverse negation queries, and introduce a dynamic visual focus module for verification. Our evaluation reveals that advanced RS MLLMs struggle with negation, exhibiting hallucinations and substantial performance degradation. To close this gap, we propose NeFo, a novel test-time learning method that explicitly incorporates the logical role of negation into the model optimization. Remarkably, using about 5\% unlabeled test samples, NeFo significantly improves the negation understanding of models and shows strong generalization to unseen tasks.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper introduces RS-Neg, the first benchmark for negation comprehension in remote sensing MLLMs spanning region- to scene-level tasks. It uses an automated LLM-based pipeline to synthesize negation queries paired with a dynamic visual focus module for verification, demonstrates that existing advanced RS MLLMs exhibit hallucinations and substantial performance drops on negation, and proposes NeFo, a test-time learning method that incorporates negation's logical role to improve understanding and generalize to unseen tasks using roughly 5% unlabeled test samples.

Significance. If the benchmark construction and NeFo gains prove robust, the work addresses a practically important gap in MLLM reliability for remote-sensing applications that require explicit reasoning about absence (e.g., evacuation routing). The automated synthesis approach and low-data test-time adaptation are efficient and potentially reusable; the empirical framing with a new benchmark constitutes a clear contribution.

major comments (2)
  1. [Benchmark construction] Benchmark construction section: the automated data-generation pipeline that relies on LLMs to synthesize negation queries and the dynamic visual focus module lack reported validation against human judgments or ground-truth RS imagery; without such checks the benchmark's representativeness remains unverified and is load-bearing for all downstream claims of model failure.
  2. [Evaluation] Evaluation section: the manuscript does not specify the exact metrics or protocols used to quantify hallucinations and performance degradation, nor does it report statistical significance or variance for the NeFo improvements; these omissions prevent assessment of whether the reported gains are reliable.
minor comments (1)
  1. [Abstract / Introduction] The abstract and introduction could more explicitly list the concrete tasks included in RS-Neg and the precise definition of "unseen tasks" used for the generalization experiments.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive feedback. The two major comments highlight important aspects of benchmark validation and evaluation rigor that we will address through targeted revisions to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Benchmark construction] Benchmark construction section: the automated data-generation pipeline that relies on LLMs to synthesize negation queries and the dynamic visual focus module lack reported validation against human judgments or ground-truth RS imagery; without such checks the benchmark's representativeness remains unverified and is load-bearing for all downstream claims of model failure.

    Authors: We agree that explicit validation of the automated pipeline is necessary to confirm representativeness. The dynamic visual focus module performs automated verification by cross-referencing synthesized queries against image content, but this is internal to the pipeline. In the revision we will add a new subsection reporting human evaluation results on a random sample of 200 generated queries (with inter-annotator agreement), comparing them to ground-truth RS imagery annotations where available. This will directly address the concern and support the downstream claims. revision: yes

  2. Referee: [Evaluation] Evaluation section: the manuscript does not specify the exact metrics or protocols used to quantify hallucinations and performance degradation, nor does it report statistical significance or variance for the NeFo improvements; these omissions prevent assessment of whether the reported gains are reliable.

    Authors: We acknowledge the omission. The current manuscript describes hallucination as incorrect affirmative responses to negation queries and performance degradation via accuracy drops, but does not detail the precise counting protocol or error taxonomy. In revision we will add an explicit subsection defining the metrics (including hallucination rate as the proportion of affirmative answers on negation queries), the evaluation protocol, and results with mean, standard deviation, and statistical significance tests (paired t-tests or Wilcoxon) across three random seeds for all NeFo experiments. This will allow readers to assess reliability of the reported gains. revision: yes

Circularity Check

0 steps flagged

No significant circularity; empirical benchmark and test-time method are self-contained

full rationale

The paper presents an empirical study: it constructs the RS-Neg benchmark via an LLM-based synthesis pipeline, evaluates existing RS MLLMs on negation tasks, and introduces NeFo as a test-time optimization technique that incorporates negation logic. No equations, fitted parameters, or derivations are described that reduce by construction to the inputs. No self-citation chains or uniqueness theorems are invoked as load-bearing premises. The central claims rest on direct performance measurements against the introduced benchmark and generalization tests, which are externally falsifiable and do not collapse into self-definition or renaming of known results.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract provides no information on free parameters, axioms, or invented entities.

pith-pipeline@v0.9.1-grok · 5713 in / 1306 out tokens · 31449 ms · 2026-06-26T18:06:47.221717+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

44 extracted references · 1 canonical work pages

  1. [1]

    In: Proceedings of the Computer Vision and Pattern Recognition Conference

    Alhamoud, K., Alshammari, S., Tian, Y., Li, G., Torr, P.H., Kim, Y., Ghassemi, M.: Vision-language models do not understand negation. In: Proceedings of the Computer Vision and Pattern Recognition Conference. pp. 29612–29622 (2025)

    2025

  2. [2]

    5-vl technical report

    Bai, S., Chen, K., Liu, X., Wang, J., Ge, W., Song, S., Dang, K., Wang, P., Wang, S., Tang, J., et al.: Qwen2. 5-vl technical report. arXiv preprint arXiv:2502.13923 (2025)

    Pith/arXiv arXiv 2025

  3. [3]

    Remote Sensing16(9), 1477 (2024)

    Bazi, Y., Bashmal, L., Al Rahhal, M.M., Ricci, R., Melgani, F.: Rs-llava: A large vision-language model for joint captioning and question answering in remote sens- ing imagery. Remote Sensing16(9), 1477 (2024)

    2024

  4. [4]

    arXiv preprint arXiv:2412.05271 (2024)

    Chen, Z., Wang, W., Cao, Y., Liu, Y., Gao, Z., Cui, E., Zhu, J., Ye, S., Tian, H., Liu, Z., et al.: Expanding performance boundaries of open-source multimodal models with model, data, and test-time scaling. arXiv preprint arXiv:2412.05271 (2024)

    Pith/arXiv arXiv 2024

  5. [5]

    IEEE Transactions on Geo- science and Remote Sensing60, 1–19 (2022)

    Cheng, Q., Huang, H., Xu, Y., Zhou, Y., Li, H., Wang, Z.: Nwpu-captions dataset and mlca-net for remote sensing image captioning. IEEE Transactions on Geo- science and Remote Sensing60, 1–19 (2022)

    2022

  6. [6]

    Advances in neural information processing systems36, 49250–49267 (2023)

    Dai, W., Li, J., Li, D., Tiong, A., Zhao, J., Wang, W., Li, B., Fung, P.N., Hoi, S.: Instructblip: Towards general-purpose vision-language models with instruction tuning. Advances in neural information processing systems36, 49250–49267 (2023)

    2023

  7. [7]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Danish, M., Munir, M.A., Shah, S.R.A., Kuckreja, K., Khan, F.S., Fraccaro, P., Lacoste, A., Khan, S.: Geobench-vlm: Benchmarking vision-language models for geospatial tasks. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 7132–7142 (2025)

    2025

  8. [8]

    Oxford University Press (2020)

    Déprez, V., Déprez, V.M., Espinal, M.T., i Farré, M.T.E.: The Oxford handbook of negation. Oxford University Press (2020)

    2020

  9. [9]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Guo, X., Lao, J., Dang, B., Zhang, Y., Yu, L., Ru, L., Zhong, L., Huang, Z., Wu, K., Hu, D., et al.: Skysense: A multi-modal remote sensing foundation model towards universal interpretation for earth observation imagery. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 27672–27683 (2024)

    2024

  10. [10]

    IEEE Transactions on Pattern Analysis and Machine In- telligence (2026)

    Han, H., Wang, A.J., Liu, F., Zhu, J.: Negation-aware test-time adaptation for vision-language models. IEEE Transactions on Pattern Analysis and Machine In- telligence (2026)

    2026

  11. [11]

    Journal of Pragmatics38(7), 1015–1032 (2006)

    Hasson, U., Glucksberg, S.: Does understanding negation entail affirmation?: An examination of negated metaphors. Journal of Pragmatics38(7), 1015–1032 (2006)

    2006

  12. [12]

    David Hume series, CSLI (2001),https: //books.google.com.sg/books?id=hBFtAAAAIAAJ

    Horn, L.: A Natural History of Negation. David Hume series, CSLI (2001),https: //books.google.com.sg/books?id=hBFtAAAAIAAJ

    2001

  13. [13]

    In: Singh, A., Fazel, M., Hsu, D., Lacoste-Julien, S., Berkenkamp, F., Maharaj, T., Wagstaff, K., Zhu, J

    Hu, J., Zhang, Z., Chen, G., Wen, X., Shuai, C., Luo, W., Xiao, B., Li, Y., Tan, M.: Test-time learning for large language models. In: Singh, A., Fazel, M., Hsu, D., Lacoste-Julien, S., Berkenkamp, F., Maharaj, T., Wagstaff, K., Zhu, J. (eds.) 16 H. Han et al. Proceedings of the 42nd International Conference on Machine Learning. Proceed- ings of Machine L...

    2025

  14. [14]

    In: International conference on machine learning

    Jia, C., Yang, Y., Xia, Y., Chen, Y.T., Parekh, Z., Pham, H., Le, Q., Sung, Y.H., Li, Z., Duerig, T.: Scaling up visual and vision-language representation learning with noisy text supervision. In: International conference on machine learning. pp. 4904–4916. PMLR (2021)

    2021

  15. [15]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Kuckreja, K., Danish, M.S., Naseer, M., Das, A., Khan, S., Khan, F.S.: Geochat: Grounded large vision-language model for remote sensing. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 27831– 27840 (2024)

    2024

  16. [16]

    In: Proceedings of the Computer Vision and Pattern Recognition Conference

    Li, G., Xu, J., Zhao, Y., Peng, Y.: Dyfo: A training-free dynamic focus visual search for enhancing lmms in fine-grained visual understanding. In: Proceedings of the Computer Vision and Pattern Recognition Conference. pp. 9098–9108 (2025)

    2025

  17. [17]

    In: IGARSS 2025-2025 IEEE International Geoscience and Remote Sensing Symposium

    Li, Q., He, X., Shu, X., Yu, Y., Chen, D., Chen, Y., Yang, X.: A simple aerial detection baseline of multimodal language models. In: IGARSS 2025-2025 IEEE International Geoscience and Remote Sensing Symposium. pp. 6833–6837. IEEE (2025)

    2025

  18. [18]

    Remote Sensing18(2), 222 (2026)

    Li, Q., Ma, S., Luo, J., Yu, Y., Zhou, Y., Wang, F., Lu, X., Wang, X., He, X., Chen, Y., et al.: Co-training vision-language models for remote sensing multi-task learning. Remote Sensing18(2), 222 (2026)

    2026

  19. [19]

    arXiv preprint arXiv:2601.04118 (2026)

    Li, W., Xiang, X., Wen, Z., Zhou, G., Niu, B., Wang, F., Huang, L., Wang, Q., Hu, Y.: Georeason: Aligning thinking and answering in remote sensing vision- language models via logical consistency reinforcement learning. arXiv preprint arXiv:2601.04118 (2026)

    arXiv 2026

  20. [20]

    Advances in Neural Information Processing Systems37, 3229–3242 (2024)

    Li, X., Ding, J., Elhoseiny, M.: Vrsbench: A versatile vision-language benchmark dataset for remote sensing image understanding. Advances in Neural Information Processing Systems37, 3229–3242 (2024)

    2024

  21. [21]

    Advances in neural information processing systems36, 34892–34916 (2023)

    Liu, H., Li, C., Wu, Q., Lee, Y.J.: Visual instruction tuning. Advances in neural information processing systems36, 34892–34916 (2023)

    2023

  22. [22]

    In: European conference on computer vision

    Liu, S., Zeng, Z., Ren, T., Li, F., Zhang, H., Yang, J., Jiang, Q., Li, C., Yang, J., Su, H., et al.: Grounding dino: Marrying dino with grounded pre-training for open-set object detection. In: European conference on computer vision. pp. 38–55. Springer (2024)

    2024

  23. [23]

    arXiv preprint arXiv:2403.05525 (2024)

    Lu, H., Liu, W., Zhang, B., Wang, B., Dong, K., Liu, B., Sun, J., Ren, T., Li, Z., Yang, H., et al.: Deepseek-vl: towards real-world vision-language understanding. arXiv preprint arXiv:2403.05525 (2024)

    Pith/arXiv arXiv 2024

  24. [24]

    IEEE Transactions on Geoscience and Remote Sensing 56(4), 2183–2195 (2017)

    Lu, X., Wang, B., Zheng, X., Li, X.: Exploring models and data for remote sensing image caption generation. IEEE Transactions on Geoscience and Remote Sensing 56(4), 2183–2195 (2017)

    2017

  25. [25]

    Ma, Z., Hong, J., Gul, M.O., Gandhi, M., Gao, I., Krishna, R.: Crepe: Can vision-language foundation models reason compositionally? In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 10910– 10921 (2023)

    2023

  26. [26]

    In: European Conference on Computer Vision

    Muhtar, D., Li, Z., Gu, F., Zhang, X., Xiao, P.: Lhrs-bot: Empowering remote sens- ing with vgi-enhanced large multimodal language model. In: European Conference on Computer Vision. pp. 440–457. Springer (2024)

    2024

  27. [27]

    In: European Conference on Computer Vision

    Nedungadi, V., Kariryaa, A., Oehmcke, S., Belongie, S., Igel, C., Lang, N.: Mmearth: Exploring multi-modal pretext tasks for geospatial representation learn- ing. In: European Conference on Computer Vision. pp. 164–182. Springer (2024) RS-Neg & NeFo 17

    2024

  28. [28]

    In: The Eleventh International Conference on Learning Representations (2023)

    Niu, S., Wu, J., Zhang, Y., Wen, Z., Chen, Y., Zhao, P., Tan, M.: Towards sta- ble test-time adaptation in dynamic wild world. In: The Eleventh International Conference on Learning Representations (2023)

    2023

  29. [29]

    arXiv preprint arXiv:2501.10913 (2025)

    Park, J., Lee, J., Song, J., Yu, S., Jung, D., Yoon, S.: Know" no”better: A data-driven approach for enhancing negation awareness in clip. arXiv preprint arXiv:2501.10913 (2025)

    arXiv 2025

  30. [30]

    In: 2016 International conference on computer, information and telecommunication systems (Cits)

    Qu, B., Li, X., Tao, D., Lu, X.: Deep semantic understanding of high resolution remote sensing image. In: 2016 International conference on computer, information and telecommunication systems (Cits). pp. 1–5. IEEE (2016)

    2016

  31. [31]

    In: International conference on machine learning

    Radford, A., Kim, J.W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al.: Learning transferable visual models from natural language supervision. In: International conference on machine learning. pp. 8748–8763. PMLR (2021)

    2021

  32. [32]

    IEEE Access9, 89644–89654 (2021)

    Rahnemoonfar, M., Chowdhury, T., Sarkar, A., Varshney, D., Yari, M., Murphy, R.R.: Floodnet: A high resolution aerial imagery dataset for post flood scene un- derstanding. IEEE Access9, 89644–89654 (2021)

    2021

  33. [33]

    arXiv preprint arXiv:2511.22396 (2025)

    Shao, R., Li, Z., Zhang, Z., Xu, L., He, X., Yuan, H., He, B., Dai, Y., Yan, Y., Chen, Y., et al.: Asking like socrates: Socrates helps vlms understand remote sensing images. arXiv preprint arXiv:2511.22396 (2025)

    Pith/arXiv arXiv 2025

  34. [34]

    no" to say

    Singh, J., Shrivastava, I., Vatsa, M., Singh, R., Bharati, A.: Learn" no" to say" yes" better: Improving vision-language models via negations. arXiv preprint arXiv:2403.20312 (2024)

    arXiv 2024

  35. [35]

    arXiv preprint arXiv:2006.10726 (2020)

    Wang, D., Shelhamer, E., Liu, S., Olshausen, B., Darrell, T.: Tent: Fully test-time adaptation by entropy minimization. arXiv preprint arXiv:2006.10726 (2020)

    Pith/arXiv arXiv 2006

  36. [36]

    arXiv preprint arXiv:2409.12191 (2024)

    Wang, P., Bai, S., Tan, S., Wang, S., Fan, Z., Bai, J., Chen, K., Liu, X., Wang, J., Ge, W., et al.: Qwen2-vl: Enhancing vision-language model’s perception of the world at any resolution. arXiv preprint arXiv:2409.12191 (2024)

    Pith/arXiv arXiv 2024

  37. [37]

    IEEE Transactions on Geoscience and Remote Sensing55(7), 3965–3981 (2017)

    Xia, G.S., Hu, J., Hu, F., Shi, B., Bai, X., Zhong, Y., Zhang, L., Lu, X.: Aid: A benchmark data set for performance evaluation of aerial scene classification. IEEE Transactions on Geoscience and Remote Sensing55(7), 3965–3981 (2017)

    2017

  38. [38]

    In: Proceedings of the 18th SIGSPATIAL international conference on advances in geographic information systems

    Yang, Y., Newsam, S.: Bag-of-visual-words and spatial extensions for land-use classification. In: Proceedings of the 18th SIGSPATIAL international conference on advances in geographic information systems. pp. 270–279 (2010)

    2010

  39. [39]

    Yuksekgonul, M., Bianchi, F., Kalluri, P., Jurafsky, D., Zou, J.: When and why vision-language models behave like bags-of-words, and what to do about it? arXiv preprint arXiv:2210.01936 (2022)

    arXiv 2022

  40. [40]

    IEEE Transactions on Geoscience and Remote Sensing62, 1–15 (2024).https://doi.org/10.1109/TGRS.2024.3363057

    Zhang, Y., Ye, M., Zhu, G., Liu, Y., Guo, P., Yan, J.: Ffca-yolo for small object detection in remote sensing images. IEEE Transactions on Geoscience and Remote Sensing62, 1–15 (2024).https://doi.org/10.1109/TGRS.2024.3363057

    work page doi:10.1109/tgrs.2024.3363057 2024

  41. [41]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Zhang, Y., Ru, L., Wu, K., Yu, L., Liang, L., Li, Y., Chen, J.: Skysense v2: A unified foundation model for multi-modal remote sensing. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 9136–9146 (2025)

    2025

  42. [42]

    Zhang, Y., Su, Y., Liu, Y., Yeung-Levy, S.: Negvqa: Can vision language models understand negation? arXiv preprint arXiv:2505.22946 (2025)

    arXiv 2025

  43. [43]

    arXiv preprint arXiv:2501.19017 (2025)

    Zhu, B., Qi, H., Gui, Y., Chen, J., Ngo, C.W., Lim, E.P.: Calling a spade a heart: Gaslighting multimodal large language models via negation. arXiv preprint arXiv:2501.19017 (2025)

    arXiv 2025

  44. [44]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Zhu, Q., Lao, J., Ji, D., Luo, J., Wu, K., Zhang, Y., Ru, L., Wang, J., Chen, J., Yang, M., et al.: Skysense-o: Towards open-world remote sensing interpretation with vision-centric visual-language modeling. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 14733–14744 (2025)

    2025