{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,18]],"date-time":"2026-03-18T10:02:33Z","timestamp":1773828153662,"version":"3.50.1"},"reference-count":37,"publisher":"MDPI AG","issue":"3","license":[{"start":{"date-parts":[[2026,3,18]],"date-time":"2026-03-18T00:00:00Z","timestamp":1773792000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Information"],"abstract":"<jats:p>Large language models are increasingly used as automated graders, yet their reliability under answer-side manipulation and their behavior in multi-model panels remain insufficiently understood. This paper introduces EvalHack, a matrix benchmark in which a fixed committee of four LLMs grades university-level machine learning exam answers under a strict integer-only contract (0\u201310) grounded in instructor-authored rubric artifacts. The dataset comprises 100 students answering 10 short, open-ended items (1000 answers). For each answer, the evaluation includes a clean version and two content-preserving adversarial variants that operate only on the student text: A1, a visible coercive suffix appended to the answer, and A2, a stealth variant that uses Unicode control characters (e.g., zero-width and bidirectional marks) to embed an instruction. EvalHack instruments the full grading pipeline, recording item-level member scores, the committee aggregate, within-panel disagreement, and discrepancies to human grades. Empirically, answer-side edits induce systematic score inflation and stronger top-end concentration, with edited answers clustering near the upper end of the scale. Within-panel disagreement, measured as the range between the highest and lowest member score, varies across conditions, with median Consistency Spread values of 3.0 (clean), 2.0 (A1), and 6.0 (A2). Compared to human graders, the panel is more lenient on average (MAE = 1.897; bias human \u2212 panel = \u22121.345). Finally, grouping items by disagreement shows that low-disagreement items exhibit smaller human-panel errors, indicating that within-panel spread can serve as a practical uncertainty signal for routing difficult answers to human review or to larger\/more specialized panels.<\/jats:p>","DOI":"10.3390\/info17030297","type":"journal-article","created":{"date-parts":[[2026,3,18]],"date-time":"2026-03-18T07:35:08Z","timestamp":1773819308000},"page":"297","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["EvalHack: Answer-Side Prompt Injection for Probing LLM Exam-Grading Panel Stability"],"prefix":"10.3390","volume":"17","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-1849-3072","authenticated-orcid":false,"given":"Catalin","family":"Anghel","sequence":"first","affiliation":[{"name":"Department of Computer Science and Information Technology, \u201cDun\u0103rea de Jos\u201d University of Galati, \u0218tiin\u021bei St. 2, 800146 Galati, Romania"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0006-9970-2556","authenticated-orcid":false,"given":"Marian Viorel","family":"Craciun","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Information Technology, \u201cDun\u0103rea de Jos\u201d University of Galati, \u0218tiin\u021bei St. 2, 800146 Galati, Romania"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0935-4713","authenticated-orcid":false,"given":"Adina","family":"Cocu","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Information Technology, \u201cDun\u0103rea de Jos\u201d University of Galati, \u0218tiin\u021bei St. 2, 800146 Galati, Romania"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0008-2537-6713","authenticated-orcid":false,"given":"Andreea Alexandra","family":"Anghel","sequence":"additional","affiliation":[{"name":"Faculty of Automation, Computer Science, Electrical and Electronic Engineering, \u201cDun\u0103rea de Jos\u201d University of Galati, 800008 Galati, Romania"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0002-5261-2060","authenticated-orcid":false,"given":"Antonio Stefan","family":"Balau","sequence":"additional","affiliation":[{"name":"Faculty of Automation, Computer Science, Electrical and Electronic Engineering, \u201cDun\u0103rea de Jos\u201d University of Galati, 800008 Galati, Romania"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1317-6676","authenticated-orcid":false,"given":"Adrian","family":"Istrate","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Information Technology, \u201cDun\u0103rea de Jos\u201d University of Galati, \u0218tiin\u021bei St. 2, 800146 Galati, Romania"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-6669-1398","authenticated-orcid":false,"given":"Aurelian-Dumitrache","family":"Anghele","sequence":"additional","affiliation":[{"name":"Department of General Surgery, Faculty of Medicine and Pharmacy, \u201cDun\u0103rea de Jos\u201d University of Galati, 47 Str. Domneasc\u0103, 800201 Galati, Romania"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1968","published-online":{"date-parts":[[2026,3,18]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","unstructured":"Liusie, A., Manakul, P., and Gales, M.J.F. (2024, January 17\u201322). LLM Comparative Assessment: Zero-shot NLG Evaluation through Pairwise Comparisons using Large Language Models. Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (EACL), St. Julians, Malta.","DOI":"10.18653\/v1\/2024.eacl-long.8"},{"key":"ref_2","unstructured":"Liu, Y., Zhou, H., Guo, Z., Shareghi, E., Vuli\u0107, I., Korhonen, A., and Collier, N. (2024, January 7\u20139). Aligning with Human Judgement: The Role of Pairwise Preference in Large Language Model Evaluators. Proceedings of the COLM 2024\u2014Conference on Language Modeling, Philadelphia, PA, USA."},{"key":"ref_3","doi-asserted-by":"crossref","unstructured":"Li, Z., Peng, B., He, P., and Yan, X. (2024, January 12\u201316). Evaluating the Instruction-Following Robustness of Large Language Models to Prompt Injection. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, Miami, FL, USA.","DOI":"10.18653\/v1\/2024.emnlp-main.33"},{"key":"ref_4","doi-asserted-by":"crossref","unstructured":"Moradi, M., and Samwald, M. (2021, January 7\u201311). Evaluating the Robustness of Neural Language Models to Input Perturbations. Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing (EMNLP 2021), Punta Cana, Dominican Republic.","DOI":"10.18653\/v1\/2021.emnlp-main.117"},{"key":"ref_5","unstructured":"The MITRE Corporation (2025, November 14). MITRE Adversarial Threat Landscape for AI Systems (ATLAS\u2122). Available online: https:\/\/atlas.mitre.org."},{"key":"ref_6","doi-asserted-by":"crossref","unstructured":"Kwon, H., and Pak, W. (2024). Text-Based Prompt Injection Attack Using Mathematical Functions in Modern Large Language Models. Electronics, 13.","DOI":"10.3390\/electronics13245008"},{"key":"ref_7","doi-asserted-by":"crossref","unstructured":"Hashemi, H., Eisner, J., Rosset, C., Van Durme, B., and Kedzie, C. (2024, January 11\u201316). LLM-RUBRIC: A Multidimensional, Calibrated Approach to Automated Evaluation of Natural Language Texts. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (ACL), Bangkok, Thailand.","DOI":"10.18653\/v1\/2024.acl-long.745"},{"key":"ref_8","unstructured":"Liang, P., Bommasani, R., Lee, T., Tsipras, D., Soylu, D., Yasunaga, M., Zhang, Y., Narayanan, D., Wu, Y., and Kumar, A. (2022). Holistic Evaluation of Language Models. arXiv."},{"key":"ref_9","doi-asserted-by":"crossref","unstructured":"Li, Z., Wang, C., Ma, P., Wu, D., Wang, S., Gao, C., and Liu, Y. (2024, January 12\u201316). Split and Merge: Aligning Position Biases in LLM-based Evaluators. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP), Miami, FL, USA.","DOI":"10.18653\/v1\/2024.emnlp-main.621"},{"key":"ref_10","doi-asserted-by":"crossref","unstructured":"Hu, Z., Song, L., Zhang, J., Xiao, Z., Wang, T., Chen, Z., Yuan, N.J., Lian, J., Ding, K., and Xiong, H. (2024). Explaining Length Bias in LLM-based Preference Evaluations. arXiv.","DOI":"10.18653\/v1\/2025.findings-emnlp.358"},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Gulyamov, S., Gulyamov, S., Rodionov, A., Khursanov, R., Mekhmonov, K., Babaev, D., and Rakhimjonov, A. (2026). Prompt Injection Attacks in Large Language Models and AI Agent Systems: A Comprehensive Review of Vulnerabilities, Attack Vectors, and Defense Mechanisms. Information, 17.","DOI":"10.3390\/info17010054"},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Cassola-Bacallao, J., Morales-Donaire, J., Hern\u00e1ndez-Montoya, P., and Keith-Norambuena, B. (2026). Benchmarking LLM-as-a-Judge Models for 5W1H Extraction Evaluation. Electronics, 15.","DOI":"10.3390\/electronics15030659"},{"key":"ref_13","doi-asserted-by":"crossref","unstructured":"Anghel, C., Anghel, A.A., Pecheanu, E., Cocu, A., and Istrate, A. (2025). Diagnosing Bias and Instability in LLM Evaluation: A Scalable Pairwise Meta-Evaluator. Information, 16.","DOI":"10.3390\/info16080652"},{"key":"ref_14","doi-asserted-by":"crossref","unstructured":"Anghel, C., Anghel, A.A., Pecheanu, E., Susnea, I., Cocu, A., and Istrate, A. (2025). Multi-Model Dialectical Evaluation of LLM Reasoning Chains: A Structured Framework with Dual Scoring Agents. Informatics, 12.","DOI":"10.3390\/informatics12030076"},{"key":"ref_15","doi-asserted-by":"crossref","unstructured":"Mendon\u00e7a, P.C., Quintal, F., and Mendon\u00e7a, F. (2025). Evaluating LLMs for Automated Scoring in Formative Assessments. Appl. Sci., 15.","DOI":"10.3390\/app15052787"},{"key":"ref_16","doi-asserted-by":"crossref","unstructured":"Garc\u00eda-Varela, F., Nussbaum, M., Mendoza, M., Mart\u00ednez-Troncoso, C., and Bekerman, Z. (2025). ChatGPT as a Stable and Fair Tool for Automated Essay Scoring. Educ. Sci., 15.","DOI":"10.3390\/educsci15080946"},{"key":"ref_17","doi-asserted-by":"crossref","unstructured":"Abishethvarman, V., Sabrina, F., and Kwan, P. (2025). Knowledge Integrity in Large Language Models: A State-of-the-Art Review. Information, 16.","DOI":"10.3390\/info16121076"},{"key":"ref_18","doi-asserted-by":"crossref","unstructured":"Zhu, K., Wang, J., Zhou, J., Wang, Z., Chen, H., Wang, Y., Yang, L., Ye, W., Gong, N.Z., and Zhang, Y. (2023). PromptRobust: Towards Evaluating the Robustness of Large Language Models on Adversarial Prompts. arXiv.","DOI":"10.1145\/3689217.3690621"},{"key":"ref_19","doi-asserted-by":"crossref","unstructured":"Bernik, A., Rado\u0161evi\u0107, D., and \u010cep, A. (2025). AI-Powered Learning: Revolutionizing Education and Automated Code Evaluation. Information, 16.","DOI":"10.3390\/info16111015"},{"key":"ref_20","doi-asserted-by":"crossref","unstructured":"Jukiewicz, M., and Wyrwa, M. (2026). Can ChatGPT Replace the Teacher in Assessment? A Review of Research on the Use of Large Language Models in Grading and Providing Feedback. Appl. Sci., 16.","DOI":"10.3390\/app16020680"},{"key":"ref_21","doi-asserted-by":"crossref","unstructured":"Seo, H., Hwang, T., Jung, J., Namgoong, H., Lee, J., and Jung, S. (2025). Large Language Models as Evaluators in Education: Verification of Feedback Consistency and Accuracy. Appl. Sci., 15.","DOI":"10.3390\/app15020671"},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Gao, M., Liu, Y., Hu, X., Wan, X., Bragg, J., and Cohan, A. (May, January 29). Re-evaluating Automatic LLM System Ranking for Alignment with Human Preference. Proceedings of the Findings of the Association for Computational Linguistics: NAACL 2025, Albuquerque, NM, USA.","DOI":"10.18653\/v1\/2025.findings-naacl.260"},{"key":"ref_23","doi-asserted-by":"crossref","unstructured":"Jeong, H., Park, C., Hong, J., Lee, H., and Choo, J. (2024). The Comparative Trap: Pairwise Comparisons Amplify Biased Preferences of LLM Evaluators. arXiv.","DOI":"10.18653\/v1\/2025.blackboxnlp-1.5"},{"key":"ref_24","doi-asserted-by":"crossref","unstructured":"Cisneros-Gonz\u00e1lez, J., Gordo-Herrera, N., Barcia-Santos, I., and S\u00e1nchez-Soriano, J. (2025). JorGPT: Instructor-Aided Grading of Programming Assignments with Large Language Models (LLMs). Future Internet, 17.","DOI":"10.3390\/fi17060265"},{"key":"ref_25","doi-asserted-by":"crossref","unstructured":"Ilieva, G., Yankova, T., Ruseva, M., and Kabaivanov, S. (2025). A Framework for Generative AI-Driven Assessment in Higher Education. Information, 16.","DOI":"10.20944\/preprints202505.0800.v1"},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"19161","DOI":"10.1007\/s10489-022-03313-w","article-title":"Detecting Textual Adversarial Examples through Text Modification on Text Classification Systems","volume":"53","author":"Kwon","year":"2023","journal-title":"Appl. Intell."},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"2485","DOI":"10.1587\/transinf.2018EDP7073","article-title":"Advanced Ensemble Adversarial Example on Unknown Deep Neural Network Classifiers","volume":"E101-D","author":"Kwon","year":"2018","journal-title":"IEICE Trans. Inf. Syst."},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"10339","DOI":"10.1007\/s11042-020-09167-z","article-title":"Classification Score Approach for Detecting Adversarial Example in Deep Neural Network","volume":"80","author":"Kwon","year":"2021","journal-title":"Multimed. Tools Appl."},{"key":"ref_29","doi-asserted-by":"crossref","first-page":"3841","DOI":"10.1109\/ACCESS.2021.3080680","article-title":"Friend-Guard TextFooler Attack on Text Classification System","volume":"13","author":"Kwon","year":"2021","journal-title":"IEEE Access"},{"key":"ref_30","doi-asserted-by":"crossref","unstructured":"Milani, A., Franzoni, V., Florindi, E., Omarbekova, A., Bekmanova, G., and Yergesh, B. (2025). When AI Is Fooled: Hidden Risks in LLM-Assisted Grading. Educ. Sci., 15.","DOI":"10.3390\/educsci15111419"},{"key":"ref_31","doi-asserted-by":"crossref","unstructured":"Seneviratne, H.M.T.W., and Manathunga, S.S. (2025). Artificial intelligence assisted automated short answer question scoring tool shows high correlation with human examiner markings. BMC Med. Educ., 25.","DOI":"10.1186\/s12909-025-07718-2"},{"key":"ref_32","doi-asserted-by":"crossref","first-page":"3813","DOI":"10.1007\/s40593-025-00517-2","article-title":"Automated Grading of Open-Ended Questions in Higher Education Using GenAI Models","volume":"35","author":"Pecuchova","year":"2025","journal-title":"Int. J. Artif. Intell. Educ."},{"key":"ref_33","doi-asserted-by":"crossref","first-page":"010126","DOI":"10.1103\/PhysRevPhysEducRes.21.010126","article-title":"Grading Explanations of Problem-Solving Process and Generating Feedback Using Large Language Models at Human-Level Accuracy","volume":"21","author":"Chen","year":"2025","journal-title":"Phys. Rev. Phys. Educ. Res."},{"key":"ref_34","doi-asserted-by":"crossref","unstructured":"Shi, J., Yuan, Z., Liu, Y., Huang, Y., Zhou, P., Sun, L., and Gong, N.Z. (2024, January 14\u201318). Optimization-based Prompt Injection Attack to LLM-as-a-Judge. Proceedings of the 2024 ACM SIGSAC Conference on Computer and Communications Security (CCS 2024), Salt Lake City, UT, USA.","DOI":"10.1145\/3658644.3690291"},{"key":"ref_35","doi-asserted-by":"crossref","first-page":"e2552099","DOI":"10.1001\/jamanetworkopen.2025.52099","article-title":"Invisible Text Injection and Peer Review by AI Models","volume":"9","author":"Choi","year":"2026","journal-title":"JAMA Netw. Open"},{"key":"ref_36","doi-asserted-by":"crossref","first-page":"254","DOI":"10.1037\/1082-989X.8.3.254","article-title":"Modern robust data analysis methods: Measures of central tendency","volume":"8","author":"Wilcox","year":"2003","journal-title":"Psychol. Methods"},{"key":"ref_37","unstructured":"Neo4j, Inc. (2026, February 28). Neo4j Graph Database Platform. Available online: https:\/\/neo4j.com\/product\/neo4j-graph-database\/."}],"container-title":["Information"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2078-2489\/17\/3\/297\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,18]],"date-time":"2026-03-18T07:49:33Z","timestamp":1773820173000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2078-2489\/17\/3\/297"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3,18]]},"references-count":37,"journal-issue":{"issue":"3","published-online":{"date-parts":[[2026,3]]}},"alternative-id":["info17030297"],"URL":"https:\/\/doi.org\/10.3390\/info17030297","relation":{},"ISSN":["2078-2489"],"issn-type":[{"value":"2078-2489","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,3,18]]}}}