{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/language-models-as-black-box-optimizers-for","title":"Language Models as Black-Box Optimizers for Vision-Language Models","arxiv_id":"2309.05950","date":"2023-09-12","proceeding":"CVPR 2024 1","authors":["Shihong Liu","Zhiqiu Lin","Samuel Yu","Ryan Lee","Tiffany Ling","Deepak Pathak","Deva Ramanan"],"abstract":"Vision-language models (VLMs) pre-trained on web-scale datasets have demonstrated remarkable capabilities on downstream tasks when fine-tuned with minimal data. However, many VLMs rely on proprietary data and are not open-source, which restricts the use of white-box approaches for fine-tuning. As such, we aim to develop a black-box approach to optimize VLMs through natural language prompts, thereby avoiding the need to access model parameters, feature embeddings, or even output logits. We propose employing chat-based LLMs to search for the best text prompt for VLMs. Specifically, we adopt an automatic hill-climbing procedure that converges to an effective prompt by evaluating the performance of current prompts and asking LLMs to refine them based on textual feedback, all within a conversational process without human-in-the-loop. In a challenging 1-shot image classification setup, our simple approach surpasses the white-box continuous prompting method (CoOp) by an average of 1.5% across 11 datasets including ImageNet. Our approach also outperforms both human-engineered and LLM-generated prompts. We highlight the advantage of conversational feedback that incorporates both positive and negative prompts, suggesting that LLMs can utilize the implicit gradient direction in textual feedback for a more efficient search. In addition, we find that the text prompts generated through our strategy are not only more interpretable but also transfer well across different VLM architectures in a black-box manner. Lastly, we apply our framework to optimize the state-of-the-art black-box VLM (DALL-E 3) for text-to-image generation, prompt inversion, and personalization.","url_abs":"https://arxiv.org/abs/2309.05950v5","url_pdf":"https://arxiv.org/pdf/2309.05950v5.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"language-models-as-black-box-optimizers-for","repo_url":"https://github.com/shihongl1998/llm-as-a-blackbox-optimizer","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"few-shot-image-classification","task_name":"Few-Shot Image Classification"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"image-generation","task_name":"Image Generation"},{"task_slug":"text-to-image-generation-1","task_name":"Text to Image Generation"},{"task_slug":"text-to-image-generation","task_name":"Text-to-Image Generation"},{"task_slug":"image-classification","task_name":"image-classification"}],"methods":[{"method_slug":"clip","method_name":"CLIP"},{"method_slug":"coop","method_name":"CoOp"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2309.05950","atlas_url":"https://app.syntology.ai/?focus=2309.05950","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2309.05950"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/shihongl1998/llm-as-a-blackbox-optimizer","reach":{"status":"ok"}}],"summary":{"ran_draft_wrong":1,"ran":6,"unverified":4},"by_repo_kind":{"official":{"samples":11,"ran":7,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":11,"samples":[{"code_sha256_prefix":"f41cb1a19b154297","entry":"encode_image","repo":"shihongl1998/llm-as-a-blackbox-optimizer","repo_kind":"official","path":"auto_prompt_image_optimization.py","file_url":"https://github.com/shihongl1998/llm-as-a-blackbox-optimizer/blob/HEAD/auto_prompt_image_optimization.py","link_basis":"harvester_set","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"f41cb1a19b154297"}},{"code_sha256_prefix":"60d87e4e28792d30","entry":"get_experiment_name","repo":"shihongl1998/llm-as-a-blackbox-optimizer","repo_kind":"official","path":"utils.py","file_url":"https://github.com/shihongl1998/llm-as-a-blackbox-optimizer/blob/HEAD/utils.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"60d87e4e28792d30"}},{"code_sha256_prefix":"5bc725a3a530c2c4","entry":"get_message","repo":"shihongl1998/llm-as-a-blackbox-optimizer","repo_kind":"official","path":"prompt_pool.py","file_url":"https://github.com/shihongl1998/llm-as-a-blackbox-optimizer/blob/HEAD/prompt_pool.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"5bc725a3a530c2c4"}},{"code_sha256_prefix":"e07e33e12b2f42d7","entry":"get_split_transformer","repo":"shihongl1998/llm-as-a-blackbox-optimizer","repo_kind":"official","path":"engine/clip/partial_model.py","file_url":"https://github.com/shihongl1998/llm-as-a-blackbox-optimizer/blob/HEAD/engine/clip/partial_model.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"e07e33e12b2f42d7"}},{"code_sha256_prefix":"219ed3eef816ba50","entry":"get_split_vit","repo":"shihongl1998/llm-as-a-blackbox-optimizer","repo_kind":"official","path":"engine/clip/partial_model.py","file_url":"https://github.com/shihongl1998/llm-as-a-blackbox-optimizer/blob/HEAD/engine/clip/partial_model.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"219ed3eef816ba50"}},{"code_sha256_prefix":"0aa8eb7ae30a6420","entry":"get_text_encoder","repo":"shihongl1998/llm-as-a-blackbox-optimizer","repo_kind":"official","path":"engine/clip/partial_model.py","file_url":"https://github.com/shihongl1998/llm-as-a-blackbox-optimizer/blob/HEAD/engine/clip/partial_model.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"0aa8eb7ae30a6420"}},{"code_sha256_prefix":"e62bfda6e308ebbc","entry":"load_templates_from_txt","repo":"shihongl1998/llm-as-a-blackbox-optimizer","repo_kind":"official","path":"load_laioncoco_templates.py","file_url":"https://github.com/shihongl1998/llm-as-a-blackbox-optimizer/blob/HEAD/load_laioncoco_templates.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"e62bfda6e308ebbc"}},{"code_sha256_prefix":"31025d8eb991afe7","entry":"ask_chatgpt_for_prompt","repo":"shihongl1998/llm-as-a-blackbox-optimizer","repo_kind":"official","path":"auto_prompt_prompt_inversion.py","file_url":"https://github.com/shihongl1998/llm-as-a-blackbox-optimizer/blob/HEAD/auto_prompt_prompt_inversion.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"31025d8eb991afe7"}},{"code_sha256_prefix":"c47aa9e5b049a11d","entry":"build_model","repo":"shihongl1998/llm-as-a-blackbox-optimizer","repo_kind":"official","path":"engine/clip/model.py","file_url":"https://github.com/shihongl1998/llm-as-a-blackbox-optimizer/blob/HEAD/engine/clip/model.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"c47aa9e5b049a11d"}},{"code_sha256_prefix":"5ca56bc26635e339","entry":"get_init_templates","repo":"shihongl1998/llm-as-a-blackbox-optimizer","repo_kind":"official","path":"utils.py","file_url":"https://github.com/shihongl1998/llm-as-a-blackbox-optimizer/blob/HEAD/utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"5ca56bc26635e339"}},{"code_sha256_prefix":"92f525cf6230e44e","entry":"read_classnames","repo":"shihongl1998/llm-as-a-blackbox-optimizer","repo_kind":"official","path":"engine/datasets/imagenet.py","file_url":"https://github.com/shihongl1998/llm-as-a-blackbox-optimizer/blob/HEAD/engine/datasets/imagenet.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"92f525cf6230e44e"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}