{"about":{"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.","site":"https://codewithpapers.app","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","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/paper/adversarial-glue-a-multi-task-benchmark-for","title":"Adversarial GLUE: A Multi-Task Benchmark for Robustness Evaluation of Language Models","arxiv_id":"2111.02840","date":"2021-11-04","proceeding":null,"authors":["Boxin Wang","Chejian Xu","Shuohang Wang","Zhe Gan","Yu Cheng","Jianfeng Gao","Ahmed Hassan Awadallah","Bo Li"],"abstract":"Large-scale pre-trained language models have achieved tremendous success across a wide range of natural language understanding (NLU) tasks, even surpassing human performance. However, recent studies reveal that the robustness of these models can be challenged by carefully crafted textual adversarial examples. While several individual datasets have been proposed to evaluate model robustness, a principled and comprehensive benchmark is still missing. In this paper, we present Adversarial GLUE (AdvGLUE), a new multi-task benchmark to quantitatively and thoroughly explore and evaluate the vulnerabilities of modern large-scale language models under various types of adversarial attacks. In particular, we systematically apply 14 textual adversarial attack methods to GLUE tasks to construct AdvGLUE, which is further validated by humans for reliable annotations. Our findings are summarized as follows. (i) Most existing adversarial attack algorithms are prone to generating invalid or ambiguous adversarial examples, with around 90% of them either changing the original semantic meanings or misleading human annotators as well. Therefore, we perform a careful filtering process to curate a high-quality benchmark. (ii) All the language models and robust training methods we tested perform poorly on AdvGLUE, with scores lagging far behind the benign accuracy. We hope our work will motivate the development of new adversarial attacks that are more stealthy and semantic-preserving, as well as new robust language models against sophisticated adversarial attacks. AdvGLUE is available at https://adversarialglue.github.io.","url_abs":"https://arxiv.org/abs/2111.02840v2","url_pdf":"https://arxiv.org/pdf/2111.02840v2.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":"adversarial-glue-a-multi-task-benchmark-for","repo_url":"https://github.com/ai-secure/adversarial-glue","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"adversarial-attack","task_name":"Adversarial Attack"},{"task_slug":"adversarial-robustness","task_name":"Adversarial Robustness"},{"task_slug":"natural-language-understanding","task_name":"Natural Language Understanding"}],"methods":[],"datasets_introduced":[{"slug":"advglue","name":"AdvGLUE","full_name":"Adversarial GLUE"}],"methods_introduced":[],"results":[{"leaderboard":"/sota/adversarial-robustness-on-advglue","task":"Adversarial Robustness","dataset":"AdvGLUE","model":"DeBERTa (single model)","rank_in_archive_order":1,"of":10,"metrics":{"Accuracy":"0.6086"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-advglue","task":"Adversarial Robustness","dataset":"AdvGLUE","model":"ALBERT (single model)","rank_in_archive_order":2,"of":10,"metrics":{"Accuracy":"0.5922"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-advglue","task":"Adversarial Robustness","dataset":"AdvGLUE","model":"T5 (single model)","rank_in_archive_order":3,"of":10,"metrics":{"Accuracy":"0.5682"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-advglue","task":"Adversarial Robustness","dataset":"AdvGLUE","model":"SMART_RoBERTa (single model)","rank_in_archive_order":4,"of":10,"metrics":{"Accuracy":"0.5371"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-advglue","task":"Adversarial Robustness","dataset":"AdvGLUE","model":"FreeLB (single model)","rank_in_archive_order":5,"of":10,"metrics":{"Accuracy":"0.5048"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-advglue","task":"Adversarial Robustness","dataset":"AdvGLUE","model":"RoBERTa (single model)","rank_in_archive_order":6,"of":10,"metrics":{"Accuracy":"0.5021"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-advglue","task":"Adversarial Robustness","dataset":"AdvGLUE","model":"InfoBERT (single model)","rank_in_archive_order":7,"of":10,"metrics":{"Accuracy":"0.4603"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-advglue","task":"Adversarial Robustness","dataset":"AdvGLUE","model":"ELECTRA (single model)","rank_in_archive_order":8,"of":10,"metrics":{"Accuracy":"0.4169"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-advglue","task":"Adversarial Robustness","dataset":"AdvGLUE","model":"BERT (single model)","rank_in_archive_order":9,"of":10,"metrics":{"Accuracy":"0.3369"},"uses_additional_data":false},{"leaderboard":"/sota/adversarial-robustness-on-advglue","task":"Adversarial Robustness","dataset":"AdvGLUE","model":"SMART_BERT (single model)","rank_in_archive_order":10,"of":10,"metrics":{"Accuracy":"0.3029"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2111.02840","atlas_url":"https://app.syntology.ai/?focus=2111.02840","mcp":null,"developers":"https://syntology.ai/developers","held":true,"reason":"arxiv_id_in_atlas","checked_against":{"atlas_ids":549449,"atlas_ids_sha256":"afd72399dafd9906e9062927f70655952385a3d75488558e1f0484e950ab09b7","read_at":"2026-10-05T01:46:21Z"},"agent_calls":[{"tool":"get_citation_path","arguments":{"paper_1":"2111.02840"},"arguments_in":{"paper_2":"another paper's arXiv id or title"},"call":"get_citation_path(paper_1=\"2111.02840\", paper_2=\"…\")"},{"tool":"get_concepts_for_paper","arguments":{"arxiv_id":"2111.02840"},"call":"get_concepts_for_paper(arxiv_id=\"2111.02840\")"}],"by_repo":[{"repo":"ai-secure/adversarial-glue","kind":"listed","samples":0,"ran":0,"instrument":0,"unverified":0}],"by_repo_is":"one row per repository in the archive's code links, Syntology's graph links or the samples; samples 0 means no sample is linked to this paper, a statement about Syntology's coverage, not about the repository: with harvested_for_other_papers present, Syntology harvested the repository for that many other papers, and without it, Syntology harvested nothing from it; placed_by_identical_code counts samples whose link records no repository, placed here because identical code was harvested from this repository for another paper; ran counts executions on a synthesized input and instrument counts failures of Syntology's instrument, not of the code"},"code_links_note":null,"arxiv_metadata":null,"syntology_extracted_results":null,"record_sha256":"9cb78cda095989477fa44b34e97423ac7fc0eb4eca6acda236166c5d129c28b4","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}