Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

MIND: A Multi-agent Framework for Zero-shot Harmful Meme Detection

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A multi-agent framework that retrieves similar unlabeled memes and debates their harmfulness lifts zero-shot detection by up to 14.7 macro-F1 points.

desk verdict Useful practical framework, but the zero-shot claim is weakened by same-benchmark reference retrieval, and the promised code isn't in the submission. read the letter →

arxiv 2507.06908 v1 pith:MKPABEWV submitted 2025-07-09 cs.CL cs.AI

Ziyan Liu , Chunxiao Fan , Haoran Lou , Yuexin Wu , Kaiwei Deng This is my paper
classification cs.CLcs.AI
keywords zero-shotharmfulmemedetectionmulti-agentframeworkretrieval-augmentedreasoningmultimodalunderstandingbidirectionalinsightderivationdebater-judgedebatelargemodelsunannotatedreferenceset
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper argues that harmful meme detection can be done with no annotated training data by having a team of multimodal agents reason over retrieved examples. The proposed framework, MIND, retrieves a few unlabeled memes similar to the target, extracts insights from them by processing the set in both forward and backward order, and then has debater agents argue for harmfulness before a judge arbitrates disagreements. On three meme benchmarks, the authors report that this pipeline lifts the macro-F1 of a 13-billion-parameter open model by 14.74, 7.70, and 13.32 points, with smaller but consistent gains on other model scales and on a closed-source model. If correct, the result matters because it offers a path to flag harmful memes during new events before annotated datasets for those events exist. The paper itself concedes that performance depends on the quality of retrieved similar memes and that the pipeline costs roughly eight times more compute than a single zero-shot call; it also contains an internal inconsistency, with the abstract promising public code while an appendix states the code is withheld pending acceptance.

What carries the argument

The machinery is the three-stage pipeline: similarity retrieval in a fused embedding space (a weighted sum of a frozen visual encoder and a frozen text encoder, scored by cosine similarity); a bidirectional insight-derivation loop in which a multimodal language agent maintains a growing set of general harmfulness rules while reading the K retrieved memes once forward and once backward; and a debater-judge decision module in which two debater agents each reason with one direction's insights and a judge agent decides when the two disagree. The load rests on the accumulated rules being general and high-level enough to transfer from unlabeled reference memes to the target meme, rather than on any weight updates or annotated examples.

What would settle it

Run MIND on the HarM test set three times, replacing the reference pool with (a) randomly relabeled memes, (b) memes from a different dataset, or (c) the top-K retrieval purged of memes that share the target meme's image template or event; if the reported macro-F1 gains over the base model collapse in any of these variants, the mechanism is reading label-correlated surface similarity rather than generalizing through the derived insights.

Watch

Extended reading notes

Core claim

The central claim is that a training-free, multi-agent pipeline can judge whether a meme is harmful by examining a handful of similar memes that have no labels. MIND runs three stages: Similar Sample Retrieval converts each meme into a multimodal embedding—a weighted sum of features from a frozen image encoder and a frozen text encoder—and keeps the top three most cosine-similar memes from an unlabeled reference set; Relevant Insight Derivation has a multimodal language agent read those three memes sequentially and accumulate a growing set of general harmfulness rules, once in forward order and once in reverse; Insight-Augmented Inference then gives each direction's rule set to a separate debater agent that issues a judgment on the target meme, and a judge agent arbitrates when the debaters disagree. On the three benchmarks, the authors report that their pipeline raises the macro-F1 of a 13-billion-parameter open model from 50.45 to 65.19 on HarM, 53.01 to 60.71 on FHM, and 55.52 to 68.84 on MAMI, and that gains appear with every backbone they tried, including a closed-source proprietary model. Ablations indicate each stage contributes, with removing the insight-derivation stage hurting most.

Load-bearing premise

The framework assumes that cosine similarity in a fixed image-plus-text embedding space selects reference memes whose content is genuinely informative about the target's harmfulness, even though the reference set comes from the same benchmark's training split and may contain near-duplicate templates; if retrieval only finds superficially similar lookalikes, the reported gains could be label leakage rather than generalizable zero-shot insight.

Editorial extensions

If this is right

  • MIND, built on a 13B open model, beats a leading closed-source model by 4.90 macro-F1 points on HarM and roughly matches a closed-source competitor on FHM.
  • The same training-free wrapper improves macro-F1 for every backbone tested (7B, 13B, and 34B open models plus a closed-source model), with gains ranging from +2.74 to +14.74 points.
  • Ablations show each stage is load-bearing: replacing retrieval with random selection costs 0.33–4.27 macro-F1 points, and removing the insight-derivation stage costs 4.69–13.26 points.
  • MIND outperforms both zero-shot and 3-shot in-context learning with the same backbone, reaching 65.19 macro-F1 on HarM versus 59.60 for 3-shot with retrieval.
  • On the FHM and MAMI benchmarks, MIND matches or beats a supervised training-based method even though it never trains on those datasets.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The same retrieve–derive–debate wrapper could plausibly transfer to other evolving multimodal moderation tasks, such as detecting harmful video frames or manipulated news images, whenever an unlabeled but recent reference pool exists.
  • Because the reference set is the same benchmark's training split, part of the reported gain may come from dataset-specific near-duplicates; a cross-benchmark reference pool (for example, HarM memes as references for MAMI) would isolate whether the derived insights transfer or the retrieval is exploiting surface similarity.
  • The framework's roughly 8x inference overhead suggests a cost-reduction path: retrieval and insight derivation could be cached per emerging meme template, making the marginal cost of moderating a new meme drop as similar memes accumulate.
  • The paper's own future directions (weighting retrieved memes by relevance, quantifying insight reliability) imply a testable variant where the judge weighs insight sets by estimated relevance instead of treating all retrieved memes uniformly.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper introduces MIND, a training-free multi-agent framework for zero-shot harmful meme detection. Given a target meme, MIND retrieves K similar memes from an unlabeled reference set using a weighted CLIP embedding (Eq. 1), derives forward and backward insights about the retrieved memes through an LMM, and then uses two debater agents plus a judge to produce the final harmful/harmless prediction. The experimental section reports large macro-F1 gains over the base LLaVA-1.5-13B model on HarM, FHM, and MAMI (+14.74, +7.70, +13.32), and consistent gains when the framework is applied to other backbones.

Significance. The proposed framework is well-motivated and has a clear architectural rationale: combining retrieval with multi-agent reasoning and debate is a plausible way to improve zero-shot multimodal classification. The paper includes useful ablations and a cross-model generalization study, and the explicit discussion of computational overhead in the limitations is a positive feature. However, the central evidence is undermined by a potential transductive leakage problem: the reference set is the same benchmark's training split, and the retrieval may return near-duplicate or same-event memes that provide label-correlated signal. If this is not controlled, the reported gains may not reflect a generalizable zero-shot mechanism, and the paper's central claim is not yet established.

major comments (3)
  1. [§3.1, §4.4 (Table 3)] The reference set Sref is the training split of the same benchmark from which Stest is drawn (§3.1). The w/o SSR ablation (§4.4) replaces retrieval with random selection from the same Sref, so it cannot distinguish similarity-based retrieval from the effect of drawing reference memes from the same event/distribution as the target. For HarM, a COVID-19 meme dataset with many template variants, the +14.74 macro-F1 improvement may be inflated by retrieving near-duplicate or same-event memes whose labels are correlated with the target's label. Please either (i) remove near-duplicates from Sref and repeat the experiments, (ii) use a reference pool drawn from a different dataset or time period, or (iii) provide an analysis of the retrieved memes showing they are not label-correlated near-duplicates.
  2. [Appendix C; §4.5 (Figure 4)] The fusion weights λv=0.8, λt=0.2 are set 'through grid search optimization within [0,1]' and K=3 is selected based on the curves in Figure 4, with both choices made on the same three benchmark test sets. This constitutes test-set-based hyperparameter selection and can overstate the framework's zero-shot performance. The authors should report the sensitivity of the main results to these parameters and either use fixed values determined before seeing the benchmarks or evaluate on a held-out validation split.
  3. [Tables 1–3, Table 4] All results are single-point estimates with no variance or significance tests. Given the small test-set sizes (e.g., HarM has 354 test memes per Table 4), a paired significance test (e.g., McNemar's) should be reported for the key comparisons in Tables 1 and 2, and ideally multiple runs should be averaged for the stochastic components (e.g., any sampling in the LMM outputs). Without these, it is difficult to confirm that the observed gains are beyond noise.
minor comments (5)
  1. [Throughout] The model name 'LLaV A' should be 'LLaVA' (the space in the name occurs in many places and appears to be a formatting artifact).
  2. [Figure 4 caption] The caption contains stray Unicode sequences such as '/uni00000015/uni00000016/...' that should be removed.
  3. [Appendix C] The prompt text says 'Your output should stricly follow'; the word 'strictly' is misspelled. Also, in the same appendix, the code-availability statement contradicts the abstract: the abstract says 'The code is available at https://github.com/destroy-lonely/MIND', but Appendix C says 'the code ... is not included in the submission' and will only be released upon acceptance. This inconsistency should be fixed.
  4. [Table 2 and §4.2] Improvements such as '+14.74%' are absolute percentage points, not relative percentages; the text should say '14.74 percentage points' to avoid ambiguity.
  5. [Appendix A, Table 4] The table lists only test-set statistics; adding the reference-set (training-split) sizes would help readers understand the retrieval pool, which is important given the central role of Sref.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: MIND's final judgments are produced by LMM inference on externally held-out test sets, and no fitted parameter, equation identity, or load-bearing self-citation forces the reported improvements.

full rationale

The claimed derivations are an inference pipeline rather than a mathematical derivation from inputs. Equations (1)-(4) fix CLIP embeddings and cosine top-K retrieval; equations (5)-(9) define forward/backward insight sets and debater/judge inference. The reported gains in Table 2 compare MIND against the base model's zero-shot predictions on the same held-out test splits, so they are empirically measurable rather than true by construction. The fusion weights lambda_v and lambda_t are stated as fixed (0.8/0.2) after grid search, but they are hyperparameters of retrieval, not fitted predictions renamed as outputs; tuning them does not make the final harmfulness label equal to a fitted value. The single citation involving a paper co-author (Huang et al., 2024, cited in Section 4.1 for the standard HarM label-merging setting) is not load-bearing: the same preprocessing is attributed to several independent prior works. The main validity caveat is experimental, not circular: the reference set Sref is the training split of the same benchmark as Stest (Section 3.1), so retrieved near-duplicates could carry label-correlated content, and the w/o SSR ablation replaces retrieval with random samples from that same distribution; this threatens the zero-shot generalization claim but does not make the prediction equivalent to its inputs by definition.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new entities. Its main free parameters are the CLIP fusion weights and the retrieval count K, both selected on the evaluated benchmarks. The framework rests on three domain assumptions: embedding similarity captures harm-relevant context, the LMM can distill reliable insights from unlabeled similar memes, and the debate mechanism improves rather than amplifies bias. The reference pool assumption (using the benchmark train split) is the least externally validated.

free parameters (3)
  • lambda_v (visual fusion weight) = 0.8
    Grid-searched in [0,1]; controls CLIP embedding blend in Eq. (1), which determines retrieved similar memes. Tuning on the same datasets without a clear validation split risks overfitting to test.
  • lambda_t (text fusion weight) = 0.2
    Companion weight to lambda_v; same grid search; directly shapes retrieval.
  • K (number of retrieved similar memes) = 3
    Chosen from Figure 4 as the peak F1 across datasets; K affects the amount of context and noise in the insights.
assumptions (4)
  • domain assumption Cosine similarity in the fused CLIP embedding space identifies memes whose content is relevant for judging the target meme's harmfulness.
    Section 3.2 uses Eq. (3) for retrieval; the ablation w/o SSR shows F1 drops, but there is no evaluation of retrieval quality or of a mismatched reference pool.
  • domain assumption An LMM can generate reliable, generalizable harmfulness insights from unlabeled similar memes, and these insights transfer to the target meme.
    Section 3.3 relies on LMMderiving to distill rules; Appendix N shows generic rules, and there is no independent check of insight quality.
  • domain assumption Forward and backward insight passes and the debater-judge debate improve decision quality rather than amplifying shared biases.
    Section 3.3.2 and 3.4 assume order-balanced and diverse reasoning; ablations show moderate drops when removed, but agents share the same backbone and may share biases.
  • domain assumption Using the benchmark training splits as the unlabeled reference pool gives a fair zero-shot evaluation.
    Section 3.1 defines Sref; retrieved memes may be near-duplicates from the same events, making the setup transductive rather than fully zero-shot and weakening the generalization claim.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MIND: A Multi-agent Framework for Zero-shot Harmful Meme Detection." pith.science (2026). https://pith.science/paper/MKPABEWV

@misc{pith2026250706908,
  author       = {Pith},
  title        = {Pith review of: MIND: A Multi-agent Framework for Zero-shot Harmful Meme Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MKPABEWV}},
  note         = {Machine review of arXiv:2507.06908}
}
read the original abstract

The rapid expansion of memes on social media has highlighted the urgent need for effective approaches to detect harmful content. However, traditional data-driven approaches struggle to detect new memes due to their evolving nature and the lack of up-to-date annotated data. To address this issue, we propose MIND, a multi-agent framework for zero-shot harmful meme detection that does not rely on annotated data. MIND implements three key strategies: 1) We retrieve similar memes from an unannotated reference set to provide contextual information. 2) We propose a bi-directional insight derivation mechanism to extract a comprehensive understanding of similar memes. 3) We then employ a multi-agent debate mechanism to ensure robust decision-making through reasoned arbitration. Extensive experiments on three meme datasets demonstrate that our proposed framework not only outperforms existing zero-shot approaches but also shows strong generalization across different model architectures and parameter scales, providing a scalable solution for harmful meme detection. The code is available at https://github.com/destroy-lonely/MIND.

Figures

Figures reproduced from arXiv: 2507.06908 by the authors.

Figure 1
Figure 1. Example of trending memes on social media, [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. An overview of our framework. MIND, for zero-shot harmful meme detection. the visual and textual modalities. We then perform the operation as in Equation 1 for all meme sam￾ples in both Stest and Sref to obtain the embeddings of all meme samples. To retrieve the most similar samples, we use cosine similarity to match the multimodal embed￾dings of the target sample and reference samples. The similarity score s is com… view at source ↗
Figure 4
Figure 4. Effect of T opK in Similar Sample Retrieval criminatory comparison between “children” and “freaks”, identifying its target toward the LGBTQ+ community shown in the visual elements. For Fig￾ure 3(c), it connects the aggressive text about being a “wife beater” with the threatening visual presen￾tation to recognize harmful gender-based messag￾ing. 2) Through deriving insights from similar memes, our proposed framework … view at source ↗
Figures from the paper (7 more)
Figure 3
Figure 3. Figure 3: Examples of correctly predicted harmful memes in (a) HarM, (b) FHM, and (c) MAMI datasets. native ability. 2) Larger K values don’t necessarily lead to better results, as they tend to incorporate less similar memes into the retrieval results, po￾tentially introducing n…
Figure 5
Figure 5. Figure 5: (a), without SSR, the framework simply interpreted it as “a satirical representation of a state￾ment made by the US President”. However, with Meme Text: Coronavirus: US President\nDonald Trump suggests\n'injecting disinfectant' to\ntreat COVID-19, 'clean the\nlungs'\nS…
Figure 6
Figure 6. Figure 6: Examples of wrongly predicted memes by our [PITH_FULL_IMAGE:figures/full_fig_p019_6.png]
Figure 7
Figure 7. Figure 7: Example of the target meme along with its similar memes from the HarM dataset. [PITH_FULL_IMAGE:figures/full_fig_p022_7.png]
Figure 8
Figure 8. Figure 8: Example of the target meme along with its similar memes from the FHM dataset. [PITH_FULL_IMAGE:figures/full_fig_p023_8.png]
Figure 9
Figure 9. Figure 9: Example of the target meme along with its similar memes from the MAMI dataset. [PITH_FULL_IMAGE:figures/full_fig_p024_9.png]
Figure 10
Figure 10. Figure 10: The prompt of Pderiving [PITH_FULL_IMAGE:figures/full_fig_p025_10.png]

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Beyond a Joke: Multi-Angle Reasoning for Detecting and Explaining Harmful Humor in Memes

    cs.AI 2026-07 conditional novelty 6.0 of 10

    MAR-12 improves humor and hate detection in memes by prompting a VLM through twelve reasoning perspectives, attention-weighting them, and generating explanations from the weighted evidence.

Reference graph

Works this paper leans on

65 extracted references · 26 canonical work pages · cited by 1 Pith paper

  1. [1]

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. Gpt-4 technical report. arXiv preprint arXiv:2303.08774

    arXiv 2023

  2. [2]

    Anas Awadalla, Irena Gao, Josh Gardner, Jack Hessel, Yusuf Hanafy, Wanrong Zhu, Kalyani Marathe, Yonatan Bitton, Samir Gadre, Shiori Sagawa, et al. 2023. Openflamingo: An open-source framework for training large autoregressive vision-language models. arXiv preprint arXiv:2308.01390

    arXiv 2023

  3. [3]

    Jinze Bai, Shuai Bai, Shusheng Yang, Shijie Wang, Sinan Tan, Peng Wang, Junyang Lin, Chang Zhou, and Jingren Zhou. 2023. Qwen-vl: A frontier large vision-language model with versatile abilities. arXiv preprint arXiv:2308.12966

    arXiv 2023

  4. [4]

    Rui Cao, Ming Shan Hee, Adriel Kuek, Wen-Haw Chong, Roy Ka-Wei Lee, and Jing Jiang. 2023. Pro-cap: Leveraging a frozen vision-language model for hateful meme detection. In Proceedings of the 31th ACM international conference on multimedia

    work page 2023

  5. [5]

    Rui Cao, Roy Ka-Wei Lee, Wen-Haw Chong, and Jing Jiang. 2022. Prompting for multimodal hateful meme classification. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 321--332

    work page 2022

  6. [6]

    Rui Cao, Roy Ka-Wei Lee, and Jing Jiang. 2024. Modularized networks for few-shot hateful meme detection. In Proceedings of the ACM on Web Conference 2024, pages 4575--4584

    work page 2024

  7. [7]

    Jun Chen, Deyao Zhu, Xiaoqian Shen, Xiang Li, Zechun Liu, Pengchuan Zhang, Raghuraman Krishnamoorthi, Vikas Chandra, Yunyang Xiong, and Mohamed Elhoseiny. 2023. Minigpt-v2: large language model as a unified interface for vision-language multi-task learning. arXiv preprint arXiv:2310.09478

    arXiv 2023

  8. [8]

    Zhe Chen, Jiannan Wu, Wenhai Wang, Weijie Su, Guo Chen, Sen Xing, Muyan Zhong, Qinglong Zhang, Xizhou Zhu, Lewei Lu, et al. 2024. Internvl: Scaling up vision foundation models and aligning for generic visual-linguistic tasks. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 24185--24198

    work page 2024

Show all 65 references
  1. [9]

    Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, and Steven Hoi. 2023. https://arxiv.org/abs/2305.06500 Instructblip: Towards general-purpose vision-language models with instruction tuning . Preprint, arXiv:2305.06500

    2023 arXiv

  2. [10]

    Mike D'Arcy, Tom Hope, Larry Birnbaum, and Doug Downey. 2024. Marg: Multi-agent review generation for scientific papers. arXiv preprint arXiv:2401.04259

    2024 arXiv

  3. [11]

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. Bert: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of NAACL-HLT, pages 4171--4186

    2019

  4. [12]

    Alexey Dosovitskiy. 2020. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929

    2020 arXiv

  5. [13]

    Yilun Du, Shuang Li, Antonio Torralba, Joshua B Tenenbaum, and Igor Mordatch. 2023. Improving factuality and reasoning in language models through multiagent debate. arXiv preprint arXiv:2305.14325

    2023 arXiv

  6. [14]

    Elisabetta Fersini, Francesca Gasparini, Giulia Rizzi, Aurora Saibene, Berta Chulvi, Paolo Rosso, Alyssa Lees, and Jeffrey Sorensen. 2022. Semeval-2022 task 5: Multimedia automatic misogyny identification. In Proceedings of the 16th International Workshop on Semantic Evaluatio...

    2022

  7. [15]

    Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat, and Mingwei Chang. 2020. Retrieval augmented language model pre-training. In International conference on machine learning, pages 3929--3938. PMLR

    2020

  8. [16]

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770--778

    2016

  9. [17]

    Ming Shan Hee, Roy Ka-Wei Lee, and Wen-Haw Chong. 2022. On explaining multimodal hateful meme detection models. In Proceedings of the ACM Web Conference 2022, pages 3651--3655

    2022

  10. [18]

    Sirui Hong, Mingchen Zhuge, Jonathan Chen, Xiawu Zheng, Yuheng Cheng, Jinlin Wang, Ceyao Zhang, Zili Wang, Steven Ka Shing Yau, Zijuan Lin, et al. 2023. Metagpt: Meta programming for multi-agent collaborative framework. In The Twelfth International Conference on Learning Repre...

    2023

  11. [19]

    Dong Huang, Qingwen Bu, Jie M Zhang, Michael Luck, and Heming Cui. 2023. Agentcoder: Multi-agent-based code generation with iterative testing and optimisation. arXiv preprint arXiv:2312.13010

    2023 arXiv

  12. [20]

    Jianzhao Huang, Hongzhan Lin, Ziyan Liu, Ziyang Luo, Guang Chen, and Jing Ma. 2024. Towards low-resource harmful meme detection with lmm agents. arXiv preprint arXiv:2411.05383

    2024 arXiv

  13. [21]

    Junhui Ji, Wei Ren, and Usman Naseem. 2023. Identifying creative harmful memes via prompt based approach. In Proceedings of the ACM Web Conference 2023, pages 3868--3872

    2023

  14. [22]

    Douwe Kiela, Suvrat Bhooshan, Hamed Firooz, Ethan Perez, and Davide Testuggine. 2019. Supervised multimodal bitransformers for classifying images and text. arXiv preprint arXiv:1909.02950

    2019 arXiv

  15. [23]

    Douwe Kiela, Hamed Firooz, Aravind Mohan, Vedanuj Goswami, Amanpreet Singh, Pratik Ringshia, and Davide Testuggine. 2020. The hateful memes challenge: Detecting hate speech in multimodal memes. Advances in neural information processing systems, 33:2611--2624

    2020

  16. [24]

    Takeshi Kojima, Shixiang Shane Gu, Machel Reid, Yutaka Matsuo, and Yusuke Iwasawa. 2022. Large language models are zero-shot reasoners. Advances in neural information processing systems, 35:22199--22213

    2022

  17. [25]

    Roy Ka-Wei Lee, Rui Cao, Ziqing Fan, Jing Jiang, and Wen-Haw Chong. 2021. Disentangling hate in online memes. In Proceedings of the 29th ACM International Conference on Multimedia, pages 5138--5147

    2021

  18. [26]

    u ttler, Mike Lewis, Wen-tau Yih, Tim Rockt \

    Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich K \"u ttler, Mike Lewis, Wen-tau Yih, Tim Rockt \"a schel, et al. 2020. Retrieval-augmented generation for knowledge-intensive nlp tasks. In Proceedings of the 34th Internat...

    2020

  19. [27]

    Tian Liang, Zhiwei He, Wenxiang Jiao, Xing Wang, Yan Wang, Rui Wang, Yujiu Yang, Shuming Shi, and Zhaopeng Tu. 2023. Encouraging divergent thinking in large language models through multi-agent debate. arXiv preprint arXiv:2305.19118

    2023 arXiv

  20. [28]

    Hongzhan Lin, Ziyang Luo, Wei Gao, Jing Ma, Bo Wang, and Ruichao Yang. 2024. Towards explainable harmful meme detection through multimodal debate between large language models. In Proceedings of the ACM on Web Conference 2024, pages 2359--2370

    2024

  21. [29]

    Hongzhan Lin, Ziyang Luo, Jing Ma, and Long Chen. 2023. Beneath the surface: Unveiling harmful memes with multimodal reasoning distilled from large language models. In The 2023 Conference on Empirical Methods in Natural Language Processing

    2023

  22. [30]

    Phillip Lippe, Nithin Holla, Shantanu Chandra, Santhosh Rajamanickam, Georgios Antoniou, Ekaterina Shutova, and Helen Yannakoudakis. 2020. A multimodal framework for the detection of hateful memes. arXiv preprint arXiv:2012.12871

    2020 arXiv

  23. [31]

    Haotian Liu, Chunyuan Li, Yuheng Li, and Yong Jae Lee. 2024. Improved baselines with visual instruction tuning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 26296--26306

    2024

  24. [32]

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. 2023 a . Visual instruction tuning. arXiv preprint arXiv:2304.08485

    2023 arXiv

  25. [33]

    Xiao Liu, Hao Yu, Hanchen Zhang, Yifan Xu, Xuanyu Lei, Hanyu Lai, Yu Gu, Hangliang Ding, Kaiwen Men, Kejuan Yang, et al. 2023 b . Agentbench: Evaluating llms as agents. In The Twelfth International Conference on Learning Representations

    2023

  26. [34]

    Aman Madaan, Niket Tandon, Prakhar Gupta, Skyler Hallinan, Luyu Gao, Sarah Wiegreffe, Uri Alon, Nouha Dziri, Shrimai Prabhumoye, Yiming Yang, et al. 2024. Self-refine: Iterative refinement with self-feedback. Advances in Neural Information Processing Systems, 36

    2024

  27. [35]

    Ning Miao, Yee Whye Teh, and Tom Rainforth. 2023. Selfcheck: Using llms to zero-shot check their own step-by-step reasoning. arXiv preprint arXiv:2308.00436

    2023 arXiv

  28. [36]

    Yao Mu, Qinglong Zhang, Mengkang Hu, Wenhai Wang, Mingyu Ding, Jun Jin, Bin Wang, Jifeng Dai, Yu Qiao, and Ping Luo. 2023. Embodiedgpt: Vision-language pre-training via embodied chain of thought. In Thirty-seventh Conference on Neural Information Processing Systems

    2023

  29. [37]

    Niklas Muennighoff. 2020. Vilio: State-of-the-art visio-linguistic models applied to hateful memes. arXiv preprint arXiv:2012.07788

    2020 arXiv

  30. [38]

    Joon Sung Park, Joseph O'Brien, Carrie Jun Cai, Meredith Ringel Morris, Percy Liang, and Michael S Bernstein. 2023. Generative agents: Interactive simulacra of human behavior. In Proceedings of the 36th annual acm symposium on user interface software and technology, pages 1--22

    2023

  31. [39]

    Shraman Pramanick, Dimitar Dimitrov, Rituparna Mukherjee, Shivam Sharma, Md Shad Akhtar, Preslav Nakov, and Tanmoy Chakraborty. 2021 a . Detecting harmful memes and their targets. arXiv preprint arXiv:2110.00413

    2021 arXiv

  32. [40]

    Shraman Pramanick, Shivam Sharma, Dimitar Dimitrov, Md Shad Akhtar, Preslav Nakov, and Tanmoy Chakraborty. 2021 b . Momenta: A multimodal framework for detecting harmful memes and their targets. arXiv preprint arXiv:2109.05184

    2021 arXiv

  33. [41]

    Chen Qian, Wei Liu, Hongzhang Liu, Nuo Chen, Yufan Dang, Jiahao Li, Cheng Yang, Weize Chen, Yusheng Su, Xin Cong, et al. 2024. Chatdev: Communicative agents for software development. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Vo...

    2024

  34. [42]

    Yiting Qu, Xinlei He, Shannon Pierson, Michael Backes, Yang Zhang, and Savvas Zannettou. 2023. On the evolution of (hateful) memes by means of multimodal contrastive learning. In 2023 IEEE Symposium on Security and Privacy (SP), pages 293--310. IEEE

    2023

  35. [43]

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. 2021. Learning transferable visual models from natural language supervision. In International conference on machine learni...

    2021

  36. [44]

    Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J Liu. 2020. Exploring the limits of transfer learning with a unified text-to-text transformer. Journal of machine learning research, 21(140):1--67

    2020

  37. [45]

    Vlad Sandulescu. 2020. Detecting hateful memes using a multimodal deep ensemble. arXiv preprint arXiv:2012.13235

    2020 arXiv

  38. [46]

    Shivam Sharma, Firoj Alam, Md Shad Akhtar, Dimitar Dimitrov, Giovanni Da San Martino, Hamed Firooz, Alon Halevy, Fabrizio Silvestri, Preslav Nakov, and Tanmoy Chakraborty. 2022. Detecting and understanding harmful memes: A survey. arXiv preprint arXiv:2205.04274

    2022 arXiv

  39. [47]

    Yongliang Shen, Kaitao Song, Xu Tan, Dongsheng Li, Weiming Lu, and Yueting Zhuang. 2023. Hugginggpt: Solving ai tasks with chatgpt and its friends in hugging face. In Thirty-seventh Conference on Neural Information Processing Systems

    2023

  40. [48]

    Karen Simonyan and Andrew Zisserman. 2014. Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556

    2014 arXiv

  41. [49]

    Chan Hee Song, Jiaman Wu, Clayton Washington, Brian M Sadler, Wei-Lun Chao, and Yu Su. 2023. Llm-planner: Few-shot grounded planning for embodied agents with large language models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 2998--3009

    2023

  42. [50]

    Haotian Sun, Yuchen Zhuang, Lingkai Kong, Bo Dai, and Chao Zhang. 2023. Adaplanner: Adaptive planning from feedback with language models. In Thirty-seventh Conference on Neural Information Processing Systems

    2023

  43. [51]

    Shardul Suryawanshi, Bharathi Raja Chakravarthi, Mihael Arcan, and Paul Buitelaar. 2020. Multimodal meme dataset (multioff) for identifying offensive content in image and text. In Proceedings of the second workshop on trolling, aggression and cyberbullying, pages 32--41

    2020

  44. [52]

    Wei Tao, Yucheng Zhou, Yanlin Wang, Wenqiang Zhang, Hongyu Zhang, and Yu Cheng. 2024. Magis: Llm-based multi-agent framework for github issue resolution. arXiv preprint arXiv:2403.17927

    2024 arXiv

  45. [53]

    Gemini Team, Petko Georgiev, Ving Ian Lei, Ryan Burnell, Libin Bai, Anmol Gulati, Garrett Tanzer, Damien Vincent, Zhufeng Pan, Shibo Wang, et al. 2024. Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context. arXiv preprint arXiv:2403.05530

    2024 arXiv

  46. [54]

    Riza Velioglu and Jewgeni Rose. 2020. Detecting hate speech in memes using multimodal deep learning approaches: Prize-winning solution to hateful memes challenge. arXiv preprint arXiv:2012.12975

    2020 arXiv

  47. [55]

    Guanzhi Wang, Yuqi Xie, Yunfan Jiang, Ajay Mandlekar, Chaowei Xiao, Yuke Zhu, Linxi Fan, and Anima Anandkumar. 2023 a . Voyager: An open-ended embodied agent with large language models. In Intrinsically-Motivated and Open-Ended Learning Workshop@ NeurIPS2023

    2023

  48. [56]

    Junyang Wang, Haiyang Xu, Haitao Jia, Xi Zhang, Ming Yan, Weizhou Shen, Ji Zhang, Fei Huang, and Jitao Sang. 2024. Mobile-agent-v2: Mobile device operation assistant with effective navigation via multi-agent collaboration. arXiv preprint arXiv:2406.01014

    2024 arXiv

  49. [57]

    Weihan Wang, Qingsong Lv, Wenmeng Yu, Wenyi Hong, Ji Qi, Yan Wang, Junhui Ji, Zhuoyi Yang, Lei Zhao, Xixuan Song, et al. 2023 b . Cogvlm: Visual expert for pretrained language models. arXiv preprint arXiv:2311.03079

    2023 arXiv

  50. [58]

    Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik R Narasimhan, and Yuan Cao. 2022. React: Synergizing reasoning and acting in language models. In The Eleventh International Conference on Learning Representations

    2022

  51. [59]

    Yifan Zeng, Yiran Wu, Xiao Zhang, Huazheng Wang, and Qingyun Wu. 2024. Autodefense: Multi-agent llm defense against jailbreak attacks. arXiv preprint arXiv:2403.04783

    2024 arXiv

  52. [60]

    Andrew Zhao, Daniel Huang, Quentin Xu, Matthieu Lin, Yong-Jin Liu, and Gao Huang. 2024. Expel: Llm agents are experiential learners. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 19632--19642

    2024

  53. [61]

    Yi Zhou, Zhenhao Chen, and Huiyuan Yang. 2021. Multimodal learning for hateful memes detection. In 2021 IEEE International Conference on Multimedia & Expo Workshops (ICMEW), pages 1--6. IEEE

    2021

  54. [62]

    Jiawen Zhu, Roy Ka-Wei Lee, and Wen Haw Chong. 2022. Multimodal zero-shot hateful meme detection. In 14th ACM Web Science Conference 2022, pages 382--389

    2022

  55. [63]

    Ron Zhu. 2020. Enhance multimodal transformer with external label and in-domain pretrain: Hateful meme challenge winning solution. arXiv preprint arXiv:2012.08290

    2020 arXiv

  56. [64]

    online" 'onlinestring :=

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list...

  57. [65]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.