{"url":"/sota/meme-classification-on-hateful-memes","task":{"name":"Meme Classification","url":"/task/meme-classification","note":null},"dataset":{"name":"Hateful Memes","url":"/dataset/hateful-memes"},"category":"Computer Vision","categories":["Computer Vision","Natural Language Processing"],"category_note":null,"description":"Meme classification refers to the task of classifying internet memes.","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["ROC-AUC","Accuracy"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"ROC-AUC":"higher","Accuracy":"higher"}},"counts":{"rows":17,"rows_with_code":16,"rows_with_paper_page":17,"rows_dated":17,"rows_using_additional_data":3},"rows":[{"rank_in_archive_order":1,"model":"RA-HMD (Qwen2-VL-7B)","metrics":{"Accuracy":"0.821","ROC-AUC":"0.911"},"uses_additional_data":false,"paper_date":"2025-02-18","paper":"/paper/improved-fine-tuning-of-large-multimodal-1","paper_url":"https://arxiv.org/abs/2502.13061v2","paper_title":"Robust Adaptation of Large Multimodal Models for Retrieval Augmented Hateful Meme Detection","code":"https://github.com/JingbiaoMei/RGCL","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":2,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":2,"model":"RA-HMD (LLaVA-1.5-7B)","metrics":{"Accuracy":"0.809","ROC-AUC":"0.897"},"uses_additional_data":false,"paper_date":"2025-02-18","paper":"/paper/improved-fine-tuning-of-large-multimodal-1","paper_url":"https://arxiv.org/abs/2502.13061v2","paper_title":"Robust Adaptation of Large Multimodal Models for Retrieval Augmented Hateful Meme Detection","code":"https://github.com/JingbiaoMei/RGCL","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":2,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":3,"model":"PaLI-X-VPD","metrics":{"ROC-AUC":"0.892"},"uses_additional_data":false,"paper_date":"2023-12-05","paper":"/paper/visual-program-distillation-distilling-tools","paper_url":"https://arxiv.org/abs/2312.03052v2","paper_title":"Visual Program Distillation: Distilling Tools and Programmatic Reasoning into Vision-Language Models","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":4,"model":"RA-HMD (Qwen2-VL-2B)","metrics":{"Accuracy":"0.791","ROC-AUC":"0.884"},"uses_additional_data":false,"paper_date":"2025-02-18","paper":"/paper/improved-fine-tuning-of-large-multimodal-1","paper_url":"https://arxiv.org/abs/2502.13061v2","paper_title":"Robust Adaptation of Large Multimodal Models for Retrieval Augmented Hateful Meme Detection","code":"https://github.com/JingbiaoMei/RGCL","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":2,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":5,"model":"RGCL (CLIP)","metrics":{"Accuracy":"0.788","ROC-AUC":"0.870"},"uses_additional_data":false,"paper_date":"2023-11-14","paper":"/paper/improving-hateful-memes-detection-via","paper_url":"https://arxiv.org/abs/2311.08110v3","paper_title":"Improving Hateful Meme Detection through Retrieval-Guided Contrastive Learning","code":"https://github.com/JingbiaoMei/RGCL","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":0,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":6,"model":"Flamingo (fine-tuned)","metrics":{"ROC-AUC":"0.866"},"uses_additional_data":false,"paper_date":"2022-04-29","paper":"/paper/flamingo-a-visual-language-model-for-few-shot-1","paper_url":"https://arxiv.org/abs/2204.14198v2","paper_title":"Flamingo: a Visual Language Model for Few-Shot Learning","code":"https://github.com/mlfoundations/open_flamingo","n_code_links":5,"syntology":{"n_ran":18,"n_unverified":6,"n_samples":24,"n_pointer_only_licence":7}},{"rank_in_archive_order":7,"model":"Hate-CLIPper - Align","metrics":{"ROC-AUC":"0.858"},"uses_additional_data":false,"paper_date":"2022-10-12","paper":"/paper/hate-clipper-multimodal-hateful-meme","paper_url":"https://arxiv.org/abs/2210.05916v3","paper_title":"Hate-CLIPper: Multimodal Hateful Meme Classification based on Cross-modal Interaction of CLIP Features","code":"https://github.com/gokulkarthik/hateclipper","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":5,"n_samples":5,"n_pointer_only_licence":0}},{"rank_in_archive_order":8,"model":"ISSUES","metrics":{"ROC-AUC":"0.855"},"uses_additional_data":false,"paper_date":"2023-10-12","paper":"/paper/mapping-memes-to-words-for-multimodal-hateful","paper_url":"https://arxiv.org/abs/2310.08368v1","paper_title":"Mapping Memes to Words for Multimodal Hateful Meme Classification","code":"https://github.com/miccunifi/issues","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":1,"n_samples":3,"n_pointer_only_licence":0}},{"rank_in_archive_order":9,"model":"Ron Zhu","metrics":{"Accuracy":"0.732","ROC-AUC":"0.845"},"uses_additional_data":true,"paper_date":"2020-12-15","paper":"/paper/enhance-multimodal-transformer-with-external","paper_url":"https://arxiv.org/abs/2012.08290v1","paper_title":"Enhance Multimodal Transformer With External Label And In-Domain Pretrain: Hateful Meme Challenge Winning Solution","code":"https://github.com/himario/hatefulmemeschallenge","n_code_links":1,"syntology":null},{"rank_in_archive_order":10,"model":"Human","metrics":{"Accuracy":"0.847","ROC-AUC":"0.8265"},"uses_additional_data":false,"paper_date":"2020-05-10","paper":"/paper/the-hateful-memes-challenge-detecting-hate","paper_url":"https://arxiv.org/abs/2005.04790v3","paper_title":"The Hateful Memes Challenge: Detecting Hate Speech in Multimodal Memes","code":"https://github.com/facebookresearch/mmf","n_code_links":4,"syntology":{"n_ran":0,"n_unverified":4,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":11,"model":"Vilio","metrics":{"Accuracy":"0.695","ROC-AUC":"0.825"},"uses_additional_data":false,"paper_date":"2020-12-14","paper":"/paper/vilio-state-of-the-art-visio-linguistic","paper_url":"https://arxiv.org/abs/2012.07788v1","paper_title":"Vilio: State-of-the-art Visio-Linguistic Models applied to Hateful Memes","code":"https://github.com/Muennighoff/vilio","n_code_links":1,"syntology":null},{"rank_in_archive_order":12,"model":"HateDetectron27","metrics":{"Accuracy":"0.765","ROC-AUC":"0.811"},"uses_additional_data":true,"paper_date":"2020-12-23","paper":"/paper/detecting-hate-speech-in-memes-using","paper_url":"https://arxiv.org/abs/2012.12975v1","paper_title":"Detecting Hate Speech in Memes Using Multimodal Deep Learning Approaches: Prize-winning solution to Hateful Memes Challenge","code":"https://github.com/rizavelioglu/hateful_memes-hate_detectron","n_code_links":1,"syntology":null},{"rank_in_archive_order":13,"model":"Pro-Cap","metrics":{"Accuracy":"0.723","ROC-AUC":"0.809"},"uses_additional_data":false,"paper_date":"2023-08-16","paper":"/paper/pro-cap-leveraging-a-frozen-vision-language","paper_url":"https://arxiv.org/abs/2308.08088v1","paper_title":"Pro-Cap: Leveraging a Frozen Vision-Language Model for Hateful Meme Detection","code":"https://github.com/social-ai-studio/pro-cap","n_code_links":2,"syntology":{"n_ran":1,"n_unverified":3,"n_samples":4,"n_pointer_only_licence":4}},{"rank_in_archive_order":14,"model":"Visual BERT COCO","metrics":{"Accuracy":"0.695","ROC-AUC":"0.754"},"uses_additional_data":false,"paper_date":"2020-05-10","paper":"/paper/the-hateful-memes-challenge-detecting-hate","paper_url":"https://arxiv.org/abs/2005.04790v3","paper_title":"The Hateful Memes Challenge: Detecting Hate Speech in Multimodal Memes","code":"https://github.com/facebookresearch/mmf","n_code_links":4,"syntology":{"n_ran":0,"n_unverified":4,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":15,"model":"SEER (RegNet10B)","metrics":{"ROC-AUC":"0.734"},"uses_additional_data":true,"paper_date":"2022-02-16","paper":"/paper/vision-models-are-more-robust-and-fair-when","paper_url":"https://arxiv.org/abs/2202.08360v2","paper_title":"Vision Models Are More Robust And Fair When Pretrained On Uncurated Images Without Supervision","code":"https://github.com/facebookresearch/vissl","n_code_links":1,"syntology":null},{"rank_in_archive_order":16,"model":"Flamingo (few-shot:32)","metrics":{"ROC-AUC":"0.700"},"uses_additional_data":false,"paper_date":"2022-04-29","paper":"/paper/flamingo-a-visual-language-model-for-few-shot-1","paper_url":"https://arxiv.org/abs/2204.14198v2","paper_title":"Flamingo: a Visual Language Model for Few-Shot Learning","code":"https://github.com/mlfoundations/open_flamingo","n_code_links":5,"syntology":{"n_ran":18,"n_unverified":6,"n_samples":24,"n_pointer_only_licence":7}},{"rank_in_archive_order":17,"model":"CLIP (zero-shot)","metrics":{"ROC-AUC":"0.661"},"uses_additional_data":false,"paper_date":"2021-02-26","paper":"/paper/learning-transferable-visual-models-from","paper_url":"https://arxiv.org/abs/2103.00020v1","paper_title":"Learning Transferable Visual Models From Natural Language Supervision","code":"https://github.com/openai/CLIP","n_code_links":82,"syntology":{"n_ran":16,"n_unverified":4,"n_samples":20,"n_pointer_only_licence":16}}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 6,795 of the 9,581 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9581,"papers_checked":6795,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":2785},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":12,"rows_with_any_sample_ran":9,"distinct_papers_with_graph_line":8,"distinct_papers_with_any_sample_ran":6,"samples_over_distinct_papers":{"n_ran":40,"n_unverified":25,"n_samples":65,"n_pointer_only_licence":27,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":62,"n_unverified":39,"n_samples":101,"n_pointer_only_licence":34,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}