{"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/data-contamination-can-cross-language","title":"Data Contamination Can Cross Language Barriers","arxiv_id":"2406.13236","date":"2024-06-19","proceeding":null,"authors":["Feng Yao","Yufan Zhuang","Zihao Sun","Sunan Xu","Animesh Kumar","Jingbo Shang"],"abstract":"The opacity in developing large language models (LLMs) is raising growing concerns about the potential contamination of public benchmarks in the pre-training data. Existing contamination detection methods are typically based on the text overlap between training and evaluation data, which can be too superficial to reflect deeper forms of contamination. In this paper, we first present a cross-lingual form of contamination that inflates LLMs' performance while evading current detection methods, deliberately injected by overfitting LLMs on the translated versions of benchmark test sets. Then, we propose generalization-based approaches to unmask such deeply concealed contamination. Specifically, we examine the LLM's performance change after modifying the original benchmark by replacing the false answer choices with correct ones from other questions. Contaminated models can hardly generalize to such easier situations, where the false choices can be \\emph{not even wrong}, as all choices are correct in their memorization. Experimental results demonstrate that cross-lingual contamination can easily fool existing detection methods, but not ours. In addition, we discuss the potential utilization of cross-lingual contamination in interpreting LLMs' working mechanisms and in post-training LLMs for enhanced multilingual capabilities. The code and dataset we use can be obtained from \\url{https://github.com/ShangDataLab/Deep-Contam}.","url_abs":"https://arxiv.org/abs/2406.13236v2","url_pdf":"https://arxiv.org/pdf/2406.13236v2.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":"data-contamination-can-cross-language","repo_url":"https://github.com/shangdatalab/deep-contam","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"memorization","task_name":"Memorization"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2406.13236","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2406.13236"}},"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/shangdatalab/deep-contam","reach":{"status":"ok"}},{"provenance":"deterministic:regex_extraction","url":"https://github.com/ShangDataLab/Deep-Contam","reach":{"status":"ok"}}],"summary":{"ran":13,"unverified":1},"by_repo_kind":{"official":{"samples":14,"ran":13,"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":14,"samples":[{"code_sha256_prefix":"316e7af195b72405","entry":"bleu","repo":"ShangDataLab/Deep-Contam","repo_kind":"official","path":"detect_method/metrics.py","file_url":"https://github.com/ShangDataLab/Deep-Contam/blob/HEAD/detect_method/metrics.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":"316e7af195b72405"}},{"code_sha256_prefix":"5158756ddf263816","entry":"doc_to_choice","repo":"ShangDataLab/Deep-Contam","repo_kind":"official","path":"detect_method/utils.py","file_url":"https://github.com/ShangDataLab/Deep-Contam/blob/HEAD/detect_method/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":"5158756ddf263816"}},{"code_sha256_prefix":"1fe48a6ad8454871","entry":"doc_to_choice","repo":"ShangDataLab/Deep-Contam","repo_kind":"official","path":"inject/run_sft_mathqa.py","file_url":"https://github.com/ShangDataLab/Deep-Contam/blob/HEAD/inject/run_sft_mathqa.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":"1fe48a6ad8454871"}},{"code_sha256_prefix":"fa0b3817bb4daaf9","entry":"extract_option_d","repo":"ShangDataLab/Deep-Contam","repo_kind":"official","path":"detect_method/metrics.py","file_url":"https://github.com/ShangDataLab/Deep-Contam/blob/HEAD/detect_method/metrics.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":"fa0b3817bb4daaf9"}},{"code_sha256_prefix":"16d87a980525188f","entry":"format_example","repo":"ShangDataLab/Deep-Contam","repo_kind":"official","path":"inject/run_sft_mmlu.py","file_url":"https://github.com/ShangDataLab/Deep-Contam/blob/HEAD/inject/run_sft_mmlu.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":"16d87a980525188f"}},{"code_sha256_prefix":"0c5fbfb8e8c72478","entry":"format_example_arc","repo":"ShangDataLab/Deep-Contam","repo_kind":"official","path":"inject/run_sft_arc.py","file_url":"https://github.com/ShangDataLab/Deep-Contam/blob/HEAD/inject/run_sft_arc.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":"0c5fbfb8e8c72478"}},{"code_sha256_prefix":"793a8b6496415457","entry":"format_subject","repo":"ShangDataLab/Deep-Contam","repo_kind":"official","path":"inject/run_sft_mmlu.py","file_url":"https://github.com/ShangDataLab/Deep-Contam/blob/HEAD/inject/run_sft_mmlu.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":"793a8b6496415457"}},{"code_sha256_prefix":"7cc112d3f99894da","entry":"formatting_prompts_func_arc","repo":"ShangDataLab/Deep-Contam","repo_kind":"official","path":"inject/run_sft_arc.py","file_url":"https://github.com/ShangDataLab/Deep-Contam/blob/HEAD/inject/run_sft_arc.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":"7cc112d3f99894da"}},{"code_sha256_prefix":"a62987de59f6b646","entry":"formatting_prompts_func_mathqa","repo":"ShangDataLab/Deep-Contam","repo_kind":"official","path":"inject/run_sft_mathqa.py","file_url":"https://github.com/ShangDataLab/Deep-Contam/blob/HEAD/inject/run_sft_mathqa.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":"a62987de59f6b646"}},{"code_sha256_prefix":"9111ab0dca945679","entry":"formatting_prompts_func_mmlu_map","repo":"ShangDataLab/Deep-Contam","repo_kind":"official","path":"inject/run_sft_mmlu.py","file_url":"https://github.com/ShangDataLab/Deep-Contam/blob/HEAD/inject/run_sft_mmlu.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":"9111ab0dca945679"}},{"code_sha256_prefix":"cd20b28fac374fd9","entry":"get_batch","repo":"ShangDataLab/Deep-Contam","repo_kind":"official","path":"translate/run_translation_arc.py","file_url":"https://github.com/ShangDataLab/Deep-Contam/blob/HEAD/translate/run_translation_arc.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":"cd20b28fac374fd9"}},{"code_sha256_prefix":"24ab0391730d3865","entry":"get_batch","repo":"ShangDataLab/Deep-Contam","repo_kind":"official","path":"translate/run_translation_mmlu.py","file_url":"https://github.com/ShangDataLab/Deep-Contam/blob/HEAD/translate/run_translation_mmlu.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":"24ab0391730d3865"}},{"code_sha256_prefix":"7300d31865db4de0","entry":"options_string_to_list","repo":"ShangDataLab/Deep-Contam","repo_kind":"official","path":"detect_method/utils.py","file_url":"https://github.com/ShangDataLab/Deep-Contam/blob/HEAD/detect_method/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":"7300d31865db4de0"}},{"code_sha256_prefix":"b712232302e61cf6","entry":"clean_llama3","repo":"ShangDataLab/Deep-Contam","repo_kind":"official","path":"detect_method/metrics.py","file_url":"https://github.com/ShangDataLab/Deep-Contam/blob/HEAD/detect_method/metrics.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":"b712232302e61cf6"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}