{"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/smalltolarge-s2l-scalable-data-selection-for","title":"SmallToLarge (S2L): Scalable Data Selection for Fine-tuning Large Language Models by Summarizing Training Trajectories of Small Models","arxiv_id":"2403.07384","date":"2024-03-12","proceeding":null,"authors":["Yu Yang","Siddhartha Mishra","Jeffrey N Chiang","Baharan Mirzasoleiman"],"abstract":"Despite the effectiveness of data selection for large language models (LLMs) during pretraining and instruction fine-tuning phases, improving data efficiency in supervised fine-tuning (SFT) for specialized domains poses significant challenges due to the complexity of fine-tuning data. To bridge this gap, we introduce an effective and scalable data selection method for SFT, SmallToLarge (S2L), which leverages training trajectories from small models to guide the data selection for larger models. We demonstrate through extensive experiments that S2L significantly improves data efficiency in SFT for mathematical problem-solving, reducing the training data to just 11% of the original MathInstruct dataset (Yue et al., 2023) to match full dataset performance while outperforming state-of-the-art data selection algorithms by an average of 4.7% across 6 in- and out-domain evaluation datasets. Remarkably, selecting only 50K data for SFT, S2L achieves a 32.7% accuracy on the most challenging MATH (Hendrycks et al., 2021) benchmark, improving Phi-2 (Li et al., 2023b) by 16.6%. In clinical text summarization on the MIMIC-III dataset (Johnson et al., 2016), S2L again outperforms training on the full dataset using only 50% of the data. Notably, S2L can perform data selection using a reference model 40x smaller than the target model, proportionally reducing the cost of data selection.","url_abs":"https://arxiv.org/abs/2403.07384v2","url_pdf":"https://arxiv.org/pdf/2403.07384v2.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":"smalltolarge-s2l-scalable-data-selection-for","repo_url":"https://github.com/bigml-cs-ucla/s2l","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"math","task_name":"Math"},{"task_slug":"mathematical-problem-solving","task_name":"Mathematical Problem-Solving"},{"task_slug":"text-summarization","task_name":"Text Summarization"}],"methods":[{"method_slug":"sft","method_name":"SFT"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2403.07384","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2403.07384"}},"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/bigml-cs-ucla/s2l","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran":4},"by_repo_kind":{"official":{"samples":4,"ran":4,"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":0,"samples":[{"code_sha256_prefix":"ea8f24a160d0f3ed","entry":"generate_and_tokenize_prompt","repo":"bigml-cs-ucla/s2l","repo_kind":"official","path":"utils.py","file_url":"https://github.com/bigml-cs-ucla/s2l/blob/HEAD/utils.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"ea8f24a160d0f3ed"}},{"code_sha256_prefix":"484ef52491c58e86","entry":"get_checkpoint_list","repo":"bigml-cs-ucla/s2l","repo_kind":"official","path":"run_distributed_trajectories.py","file_url":"https://github.com/bigml-cs-ucla/s2l/blob/HEAD/run_distributed_trajectories.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"484ef52491c58e86"}},{"code_sha256_prefix":"a17d368b7f8d0697","entry":"get_prompter","repo":"bigml-cs-ucla/s2l","repo_kind":"official","path":"utils.py","file_url":"https://github.com/bigml-cs-ucla/s2l/blob/HEAD/utils.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"a17d368b7f8d0697"}},{"code_sha256_prefix":"e8d1298ca322021e","entry":"tokenize","repo":"bigml-cs-ucla/s2l","repo_kind":"official","path":"utils.py","file_url":"https://github.com/bigml-cs-ucla/s2l/blob/HEAD/utils.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"e8d1298ca322021e"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}