{"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/bot-breaking-long-thought-processes-of-o1","title":"To Think or Not to Think: Exploring the Unthinking Vulnerability in Large Reasoning Models","arxiv_id":"2502.12202","date":"2025-02-16","proceeding":null,"authors":["Zihao Zhu","Hongbao Zhang","Ruotong Wang","Ke Xu","Siwei Lyu","Baoyuan Wu"],"abstract":"Large Reasoning Models (LRMs) are designed to solve complex tasks by generating explicit reasoning traces before producing final answers. However, we reveal a critical vulnerability in LRMs -- termed Unthinking Vulnerability -- wherein the thinking process can be bypassed by manipulating special delimiter tokens. It is empirically demonstrated to be widespread across mainstream LRMs, posing both a significant risk and potential utility, depending on how it is exploited. In this paper, we systematically investigate this vulnerability from both malicious and beneficial perspectives. On the malicious side, we introduce Breaking of Thought (BoT), a novel attack that enables adversaries to bypass the thinking process of LRMs, thereby compromising their reliability and availability. We present two variants of BoT: a training-based version that injects backdoor during the fine-tuning stage, and a training-free version based on adversarial attack during the inference stage. As a potential defense, we propose thinking recovery alignment to partially mitigate the vulnerability. On the beneficial side, we introduce Monitoring of Thought (MoT), a plug-and-play framework that allows model owners to enhance efficiency and safety. It is implemented by leveraging the same vulnerability to dynamically terminate redundant or risky reasoning through external monitoring. Extensive experiments show that BoT poses a significant threat to reasoning reliability, while MoT provides a practical solution for preventing overthinking and jailbreaking. Our findings expose an inherent flaw in current LRM architectures and underscore the need for more robust reasoning systems in the future.","url_abs":"https://arxiv.org/abs/2502.12202v2","url_pdf":"https://arxiv.org/pdf/2502.12202v2.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":"bot-breaking-long-thought-processes-of-o1","repo_url":"https://github.com/zihao-ai/bot","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"bot-breaking-long-thought-processes-of-o1","repo_url":"https://github.com/zihao-ai/unthinking_vulnerability","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"adversarial-attack","task_name":"Adversarial Attack"},{"task_slug":"backdoor-attack","task_name":"Backdoor Attack"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=2502.12202","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2502.12202"}},"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/zihao-ai/bot","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/zihao-ai/unthinking_vulnerability","reach":null}],"summary":{"ran_violates":2,"ran_draft_wrong":3},"by_repo_kind":{"official":{"samples":5,"ran":5,"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":5,"samples":[{"code_sha256_prefix":"e903f0c23562d4f5","entry":"evaluate_question","repo":"zihao-ai/unthinking_vulnerability","repo_kind":"official","path":"MoT/generate_effiency.py","file_url":"https://github.com/zihao-ai/unthinking_vulnerability/blob/HEAD/MoT/generate_effiency.py","link_basis":"first_harvest_node","language":"python","status":"ran_violates","verification_level":1,"contract_check":"VIOLATES","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"e903f0c23562d4f5"}},{"code_sha256_prefix":"547bcf7ac9669873","entry":"evaluate_thinking","repo":"zihao-ai/unthinking_vulnerability","repo_kind":"official","path":"MoT/generate_effiency.py","file_url":"https://github.com/zihao-ai/unthinking_vulnerability/blob/HEAD/MoT/generate_effiency.py","link_basis":"first_harvest_node","language":"python","status":"ran_violates","verification_level":1,"contract_check":"VIOLATES","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"547bcf7ac9669873"}},{"code_sha256_prefix":"eaf38eec3d260e03","entry":"extract_thinking_to_last_punctuation","repo":"zihao-ai/unthinking_vulnerability","repo_kind":"official","path":"MoT/generate_effiency.py","file_url":"https://github.com/zihao-ai/unthinking_vulnerability/blob/HEAD/MoT/generate_effiency.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"eaf38eec3d260e03"}},{"code_sha256_prefix":"08cfda0dbd62bd38","entry":"poison_math_problem","repo":"zihao-ai/unthinking_vulnerability","repo_kind":"official","path":"training_based_BoT/bot_sft_lora.py","file_url":"https://github.com/zihao-ai/unthinking_vulnerability/blob/HEAD/training_based_BoT/bot_sft_lora.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"08cfda0dbd62bd38"}},{"code_sha256_prefix":"874988906a52ded5","entry":"setup_experiment_dir","repo":"zihao-ai/unthinking_vulnerability","repo_kind":"official","path":"training_based_BoT/bot_sft_lora.py","file_url":"https://github.com/zihao-ai/unthinking_vulnerability/blob/HEAD/training_based_BoT/bot_sft_lora.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"874988906a52ded5"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}