{"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/let-s-synthesize-step-by-step-iterative","title":"Let's Synthesize Step by Step: Iterative Dataset Synthesis with Large Language Models by Extrapolating Errors from Small Models","arxiv_id":"2310.13671","date":"2023-10-20","proceeding":null,"authors":["Ruida Wang","Wangchunshu Zhou","Mrinmaya Sachan"],"abstract":"*Data Synthesis* is a promising way to train a small model with very little labeled data. One approach for data synthesis is to leverage the rich knowledge from large language models to synthesize pseudo training examples for small models, making it possible to achieve both data and compute efficiency at the same time. However, a key challenge in data synthesis is that the synthesized dataset often suffers from a large distributional discrepancy from the *real task* data distribution. Thus, in this paper, we propose *Synthesis Step by Step* (**S3**), a data synthesis framework that shrinks this distribution gap by iteratively extrapolating the errors made by a small model trained on the synthesized dataset on a small real-world validation dataset using a large language model. Extensive experiments on multiple NLP tasks show that our approach improves the performance of a small model by reducing the gap between the synthetic dataset and the real data, resulting in significant improvement compared to several baselines: 9.48% improvement compared to ZeroGen and 2.73% compared to GoldGen, and at most 15.17% improvement compared to the small model trained on human-annotated data.","url_abs":"https://arxiv.org/abs/2310.13671v1","url_pdf":"https://arxiv.org/pdf/2310.13671v1.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":"let-s-synthesize-step-by-step-iterative","repo_url":"https://github.com/rickyskywalker/synthesis_step-by-step_official","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"},{"task_slug":"large-language-model","task_name":"Large Language Model"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=2310.13671","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2310.13671"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+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/rickyskywalker/synthesis_step-by-step_official","reach":{"status":"ok"}}],"summary":{"ran":7,"unverified":5},"by_repo_kind":{"official":{"samples":12,"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":12,"samples":[{"code_sha256_prefix":"0816c1b41cab192e","entry":"convertToDatasetIMDB","repo":"rickyskywalker/synthesis_step-by-step_official","repo_kind":"official","path":"ModelTraining/utils.py","file_url":"https://github.com/rickyskywalker/synthesis_step-by-step_official/blob/HEAD/ModelTraining/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":"0816c1b41cab192e"}},{"code_sha256_prefix":"ae3be87eaaf3193f","entry":"convertToDatasetQNLI","repo":"rickyskywalker/synthesis_step-by-step_official","repo_kind":"official","path":"ModelTraining/utils.py","file_url":"https://github.com/rickyskywalker/synthesis_step-by-step_official/blob/HEAD/ModelTraining/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":"ae3be87eaaf3193f"}},{"code_sha256_prefix":"3e9bdb06d0bcaee0","entry":"eval","repo":"rickyskywalker/synthesis_step-by-step_official","repo_kind":"official","path":"ModelTraining/trainHelper.py","file_url":"https://github.com/rickyskywalker/synthesis_step-by-step_official/blob/HEAD/ModelTraining/trainHelper.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":"3e9bdb06d0bcaee0"}},{"code_sha256_prefix":"1cccb768f5164e2b","entry":"f1_score","repo":"rickyskywalker/synthesis_step-by-step_official","repo_kind":"official","path":"ModelTraining/trainHelperQA.py","file_url":"https://github.com/rickyskywalker/synthesis_step-by-step_official/blob/HEAD/ModelTraining/trainHelperQA.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"1cccb768f5164e2b"}},{"code_sha256_prefix":"cc9e344fde6d9d0c","entry":"predFromLogits","repo":"rickyskywalker/synthesis_step-by-step_official","repo_kind":"official","path":"ModelTraining/trainHelper.py","file_url":"https://github.com/rickyskywalker/synthesis_step-by-step_official/blob/HEAD/ModelTraining/trainHelper.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"cc9e344fde6d9d0c"}},{"code_sha256_prefix":"53794bee860f8cf6","entry":"read_dataset","repo":"rickyskywalker/synthesis_step-by-step_official","repo_kind":"official","path":"ModelTraining/utils.py","file_url":"https://github.com/rickyskywalker/synthesis_step-by-step_official/blob/HEAD/ModelTraining/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":"53794bee860f8cf6"}},{"code_sha256_prefix":"559ca9a7878d00a3","entry":"train","repo":"rickyskywalker/synthesis_step-by-step_official","repo_kind":"official","path":"ModelTraining/trainHelperQA.py","file_url":"https://github.com/rickyskywalker/synthesis_step-by-step_official/blob/HEAD/ModelTraining/trainHelperQA.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":"559ca9a7878d00a3"}},{"code_sha256_prefix":"faf11a8545fc40a3","entry":"eval","repo":"rickyskywalker/synthesis_step-by-step_official","repo_kind":"official","path":"ModelTraining/trainHelperQA.py","file_url":"https://github.com/rickyskywalker/synthesis_step-by-step_official/blob/HEAD/ModelTraining/trainHelperQA.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":"faf11a8545fc40a3"}},{"code_sha256_prefix":"c01766547414e4b9","entry":"tokenize_function","repo":"rickyskywalker/synthesis_step-by-step_official","repo_kind":"official","path":"ModelTraining/main_IMDb.py","file_url":"https://github.com/rickyskywalker/synthesis_step-by-step_official/blob/HEAD/ModelTraining/main_IMDb.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":"c01766547414e4b9"}},{"code_sha256_prefix":"62e1275724f3399a","entry":"tokenize_function_adqa","repo":"rickyskywalker/synthesis_step-by-step_official","repo_kind":"official","path":"ModelTraining/main_AdQA.py","file_url":"https://github.com/rickyskywalker/synthesis_step-by-step_official/blob/HEAD/ModelTraining/main_AdQA.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":"62e1275724f3399a"}},{"code_sha256_prefix":"af3e38c6e91d8408","entry":"tokenize_function_qnli","repo":"rickyskywalker/synthesis_step-by-step_official","repo_kind":"official","path":"ModelTraining/main_qnli.py","file_url":"https://github.com/rickyskywalker/synthesis_step-by-step_official/blob/HEAD/ModelTraining/main_qnli.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":"af3e38c6e91d8408"}},{"code_sha256_prefix":"76b0eddf34a9bdb2","entry":"train","repo":"rickyskywalker/synthesis_step-by-step_official","repo_kind":"official","path":"ModelTraining/trainHelper.py","file_url":"https://github.com/rickyskywalker/synthesis_step-by-step_official/blob/HEAD/ModelTraining/trainHelper.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":"76b0eddf34a9bdb2"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}