{"url":"/sota/prompt-engineering-on-imagenet","task":{"name":"Prompt Engineering","url":"/task/prompt-engineering","note":null},"dataset":{"name":"ImageNet","url":"/dataset/imagenet"},"category":"Computer Vision","categories":["Computer Vision","Natural Language Processing"],"category_note":null,"description":"**Prompt engineering** is the process of designing and refining the prompts used to generate text from language models, such as GPT-3 or similar models. The goal of prompt engineering is to improve the quality and relevance of the generated text by carefully crafting the prompts to elicit the desired responses from the model.\r\n\r\nPrompt engineering involves several steps, including selecting the appropriate model architecture and parameters, designing the prompt format and structure, selecting the appropriate task and training data, and fine-tuning the model using the selected prompt and data.\r\n\r\nPrompt engineering is a crucial step in the development of language models, as it can greatly influence the quality and effectiveness of the model's responses. By carefully designing and refining the prompts used to generate text, researchers and developers can improve the accuracy and relevance of the model's output, making it more useful for a wide range of applications, including chatbots, language translation, content creation, and more.","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":["Harmonic mean"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"Harmonic mean":null}},"counts":{"rows":15,"rows_with_code":15,"rows_with_paper_page":15,"rows_dated":15,"rows_using_additional_data":0},"rows":[{"rank_in_archive_order":1,"model":"PromptKD","metrics":{"Harmonic mean":"77.62"},"uses_additional_data":false,"paper_date":"2024-03-05","paper":"/paper/promptkd-unsupervised-prompt-distillation-for","paper_url":"https://arxiv.org/abs/2403.02781v5","paper_title":"PromptKD: Unsupervised Prompt Distillation for Vision-Language Models","code":"https://github.com/zhengli97/promptkd","n_code_links":1,"syntology":null},{"rank_in_archive_order":2,"model":"Customized Ensemble","metrics":{"Harmonic mean":"75.49"},"uses_additional_data":false,"paper_date":"2023-11-28","paper":"/paper/beyond-sole-strength-customized-ensembles-for","paper_url":"https://arxiv.org/abs/2311.17091v2","paper_title":"Beyond Sole Strength: Customized Ensembles for Generalized Vision-Language Models","code":"https://github.com/zhihelu/ensemble_vlm","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":2,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":3,"model":"MMRL","metrics":{"Harmonic mean":"74.45"},"uses_additional_data":false,"paper_date":"2025-03-11","paper":"/paper/mmrl-multi-modal-representation-learning-for","paper_url":"https://arxiv.org/abs/2503.08497v2","paper_title":"MMRL: Multi-Modal Representation Learning for Vision-Language Models","code":"https://github.com/yunncheng/MMRL","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":1,"n_samples":1,"n_pointer_only_licence":0}},{"rank_in_archive_order":4,"model":"MMRL++","metrics":{"Harmonic mean":"74.44"},"uses_additional_data":false,"paper_date":"2025-05-15","paper":"/paper/mmrl-parameter-efficient-and-interaction","paper_url":"https://arxiv.org/abs/2505.10088v1","paper_title":"MMRL++: Parameter-Efficient and Interaction-Aware Representation Learning for Vision-Language Models","code":"https://github.com/yunncheng/MMRL","n_code_links":1,"syntology":null},{"rank_in_archive_order":5,"model":"CoPrompt","metrics":{"Harmonic mean":"74.33"},"uses_additional_data":false,"paper_date":"2023-06-01","paper":"/paper/consistency-guided-prompt-learning-for-vision","paper_url":"https://arxiv.org/abs/2306.01195v4","paper_title":"Consistency-guided Prompt Learning for Vision-Language Models","code":"https://github.com/shuvenduroy/coprompt","n_code_links":2,"syntology":{"n_ran":3,"n_unverified":4,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":6,"model":"HPT++","metrics":{"Harmonic mean":"74.24"},"uses_additional_data":false,"paper_date":"2024-08-27","paper":"/paper/hpt-hierarchically-prompting-vision-language","paper_url":"https://arxiv.org/abs/2408.14812v1","paper_title":"HPT++: Hierarchically Prompting Vision-Language Models with Multi-Granularity Knowledge Generation and Improved Structure Modeling","code":"https://github.com/vill-lab/2024-aaai-hpt","n_code_links":2,"syntology":null},{"rank_in_archive_order":7,"model":"HPT","metrics":{"Harmonic mean":"74.17"},"uses_additional_data":false,"paper_date":"2023-12-11","paper":"/paper/learning-hierarchical-prompt-with-structured","paper_url":"https://arxiv.org/abs/2312.06323v1","paper_title":"Learning Hierarchical Prompt with Structured Linguistic Knowledge for Vision-Language Models","code":"https://github.com/vill-lab/2024-aaai-hpt","n_code_links":2,"syntology":{"n_ran":3,"n_unverified":4,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":8,"model":"ProMetaR","metrics":{"Harmonic mean":"74.09"},"uses_additional_data":false,"paper_date":"2024-04-01","paper":"/paper/prompt-learning-via-meta-regularization","paper_url":"https://arxiv.org/abs/2404.00851v1","paper_title":"Prompt Learning via Meta-Regularization","code":"https://github.com/mlvlab/prometar","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":3,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":9,"model":"MetaPrompt","metrics":{"Harmonic mean":"74.02"},"uses_additional_data":false,"paper_date":"2022-12-08","paper":"/paper/learning-domain-invariant-prompt-for-vision","paper_url":"https://arxiv.org/abs/2212.04196v2","paper_title":"Learning Domain Invariant Prompt for Vision-Language Models","code":"https://github.com/Vill-Lab/2024-TIP-MetaPrompt","n_code_links":1,"syntology":null},{"rank_in_archive_order":10,"model":"DePT","metrics":{"Harmonic mean":"74.02"},"uses_additional_data":false,"paper_date":"2023-09-14","paper":"/paper/dept-decoupled-prompt-tuning","paper_url":"https://arxiv.org/abs/2309.07439v2","paper_title":"DePT: Decoupled Prompt Tuning","code":"https://github.com/koorye/dept","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":4,"n_samples":8,"n_pointer_only_licence":8}},{"rank_in_archive_order":11,"model":"PromptSRC","metrics":{"Harmonic mean":"74.01"},"uses_additional_data":false,"paper_date":"2023-07-13","paper":"/paper/self-regulating-prompts-foundational-model","paper_url":"https://arxiv.org/abs/2307.06948v2","paper_title":"Self-regulating Prompts: Foundational Model Adaptation without Forgetting","code":"https://github.com/muzairkhattak/promptsrc","n_code_links":2,"syntology":{"n_ran":7,"n_unverified":14,"n_samples":21,"n_pointer_only_licence":0}},{"rank_in_archive_order":12,"model":"RPO","metrics":{"Harmonic mean":"74.00"},"uses_additional_data":false,"paper_date":"2023-08-29","paper":"/paper/read-only-prompt-optimization-for-vision","paper_url":"https://arxiv.org/abs/2308.14960v2","paper_title":"Read-only Prompt Optimization for Vision-Language Few-shot Learning","code":"https://github.com/mlvlab/rpo","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":2,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":13,"model":"MaPLe","metrics":{"Harmonic mean":"73.47"},"uses_additional_data":false,"paper_date":"2022-10-06","paper":"/paper/maple-multi-modal-prompt-learning","paper_url":"https://arxiv.org/abs/2210.03117v3","paper_title":"MaPLe: Multi-modal Prompt Learning","code":"https://github.com/muzairkhattak/multimodal-prompt-learning","n_code_links":3,"syntology":{"n_ran":4,"n_unverified":2,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":14,"model":"CoCoOp","metrics":{"Harmonic mean":"73.10"},"uses_additional_data":false,"paper_date":"2022-03-10","paper":"/paper/conditional-prompt-learning-for-vision","paper_url":"https://arxiv.org/abs/2203.05557v2","paper_title":"Conditional Prompt Learning for Vision-Language Models","code":"https://github.com/kaiyangzhou/coop","n_code_links":12,"syntology":{"n_ran":4,"n_unverified":2,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":15,"model":"CLIP","metrics":{"Harmonic mean":"70.22"},"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 6,264 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":6264,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":3316},"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":11,"rows_with_any_sample_ran":10,"distinct_papers_with_graph_line":11,"distinct_papers_with_any_sample_ran":10,"samples_over_distinct_papers":{"n_ran":53,"n_unverified":42,"n_samples":95,"n_pointer_only_licence":24,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":53,"n_unverified":42,"n_samples":95,"n_pointer_only_licence":24,"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"}}}