{"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/arxiv-2602-00428","title":"When Agents \"Misremember\" Collectively: Exploring the Mandela Effect in LLM-based Multi-Agent Systems","arxiv_id":"2602.00428","date":"2026-01-31","proceeding":null,"authors":["Naen Xu","Hengyu An","Shuo Shi","Jinghuai Zhang","Chunyi Zhou","Changjiang Li","Tianyu Du","Zhihui Fu","Jun Wang","Shouling Ji"],"abstract":"Recent advancements in large language models (LLMs) have significantly enhanced the capabilities of collaborative multi-agent systems, enabling them to address complex challenges. However, within these multi-agent systems, the susceptibility of agents to collective cognitive biases remains an underexplored issue. A compelling example is the Mandela effect, a phenomenon where groups collectively misremember past events as a result of false details reinforced through social influence and internalized misinformation. This vulnerability limits our understanding of memory bias in multi-agent systems and raises ethical concerns about the potential spread of misinformation. In this paper, we conduct a comprehensive study on the Mandela effect in LLM-based multi-agent systems, focusing on its existence, causing factors, and mitigation strategies. We propose MANBENCH, a novel benchmark designed to evaluate agent behaviors across four common task types that are susceptible to the Mandela effect, using five interaction protocols that vary in agent roles and memory timescales. We evaluate agents powered by several LLMs on MANBENCH to quantify the Mandela effect and analyze how different factors affect it. Moreover, we propose strategies to mitigate this effect, including prompt-level defenses (e.g., cognitive anchoring and source scrutiny) and model-level alignment-based defense, achieving an average 74.40% reduction in the Mandela effect compared to the baseline. Our findings provide valuable insights for developing more resilient and ethically aligned collaborative multi-agent systems. Code and dataset are available at https://github.com/bluedream02/Mandela-Effect.","url_abs":"https://arxiv.org/abs/2602.00428","url_pdf":"https://arxiv.org/pdf/2602.00428","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2602.00428","atlas_url":"https://app.syntology.ai/?focus=2602.00428","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2602.00428"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+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":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/bluedream02/Mandela-Effect","reach":null}],"summary":{"ran_draft_wrong":1,"unverified":11},"by_repo_kind":{"found_in_text":{"samples":12,"ran":1,"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":"1f183be69c67626e","entry":"extract_answer","repo":"bluedream02/Mandela-Effect","repo_kind":"found_in_text","path":"eval_correct.py","file_url":"https://github.com/bluedream02/Mandela-Effect/blob/HEAD/eval_correct.py","link_basis":"plan_row","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1f183be69c67626e"}},{"code_sha256_prefix":"656ab45b98459a20","entry":"add_retries","repo":"bluedream02/Mandela-Effect","repo_kind":"found_in_text","path":"utils.py","file_url":"https://github.com/bluedream02/Mandela-Effect/blob/HEAD/utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"656ab45b98459a20"}},{"code_sha256_prefix":"b7caf13a863aed04","entry":"assign_mandela_roles","repo":"bluedream02/Mandela-Effect","repo_kind":"found_in_text","path":"format_data_bbh_combined.py","file_url":"https://github.com/bluedream02/Mandela-Effect/blob/HEAD/format_data_bbh_combined.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"b7caf13a863aed04"}},{"code_sha256_prefix":"936a599ef2fc2fec","entry":"convert_outputs_to_protocol_names","repo":"bluedream02/Mandela-Effect","repo_kind":"found_in_text","path":"eval_defense.py","file_url":"https://github.com/bluedream02/Mandela-Effect/blob/HEAD/eval_defense.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"936a599ef2fc2fec"}},{"code_sha256_prefix":"0bc875d933535de0","entry":"extract_model_name","repo":"bluedream02/Mandela-Effect","repo_kind":"found_in_text","path":"analyze.py","file_url":"https://github.com/bluedream02/Mandela-Effect/blob/HEAD/analyze.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0bc875d933535de0"}},{"code_sha256_prefix":"6554f055312adbd7","entry":"extract_task_name","repo":"bluedream02/Mandela-Effect","repo_kind":"found_in_text","path":"analyze.py","file_url":"https://github.com/bluedream02/Mandela-Effect/blob/HEAD/analyze.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"6554f055312adbd7"}},{"code_sha256_prefix":"efed52c0f3e0784b","entry":"generate_agent_names","repo":"bluedream02/Mandela-Effect","repo_kind":"found_in_text","path":"format_data_bbh_combined.py","file_url":"https://github.com/bluedream02/Mandela-Effect/blob/HEAD/format_data_bbh_combined.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"efed52c0f3e0784b"}},{"code_sha256_prefix":"ce49e844dbf4bfc8","entry":"get_cached_model_and_tokenizer","repo":"bluedream02/Mandela-Effect","repo_kind":"found_in_text","path":"utils.py","file_url":"https://github.com/bluedream02/Mandela-Effect/blob/HEAD/utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"ce49e844dbf4bfc8"}},{"code_sha256_prefix":"a6ac3727d6edecb9","entry":"get_column_name","repo":"bluedream02/Mandela-Effect","repo_kind":"found_in_text","path":"analyze.py","file_url":"https://github.com/bluedream02/Mandela-Effect/blob/HEAD/analyze.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"a6ac3727d6edecb9"}},{"code_sha256_prefix":"a57cc08fbe82267c","entry":"get_correct_guidance_prompt","repo":"bluedream02/Mandela-Effect","repo_kind":"found_in_text","path":"format_data_bbh_correct.py","file_url":"https://github.com/bluedream02/Mandela-Effect/blob/HEAD/format_data_bbh_correct.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"a57cc08fbe82267c"}},{"code_sha256_prefix":"946645c7111d72d8","entry":"get_enriching_guidance_prompt","repo":"bluedream02/Mandela-Effect","repo_kind":"found_in_text","path":"format_data_bbh_correct.py","file_url":"https://github.com/bluedream02/Mandela-Effect/blob/HEAD/format_data_bbh_correct.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"946645c7111d72d8"}},{"code_sha256_prefix":"f993de31d7dcd5dc","entry":"get_role_specific_prompt","repo":"bluedream02/Mandela-Effect","repo_kind":"found_in_text","path":"format_data_bbh_combined.py","file_url":"https://github.com/bluedream02/Mandela-Effect/blob/HEAD/format_data_bbh_combined.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"f993de31d7dcd5dc"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.CL","source":"arxiv_2026.jsonl"},"syntology_extracted_results":null}