{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:6LHSGKTTQBKCEIUU3LAJGF7IQY","short_pith_number":"pith:6LHSGKTT","schema_version":"1.0","canonical_sha256":"f2cf232a738054222294dac09317e886372e61d304637af86f7764f89084f9a6","source":{"kind":"arxiv","id":"2603.14184","version":2},"attestation_state":"computed","paper":{"title":"Deeper Thought, Weaker Aim: Understanding and Mitigating Perceptual Impairment during Reasoning in Multimodal Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Jing Lei, Lu Hou, Ruiying Peng, Xiaohui Li, Xueyu Wu, Yuanzheng Ma","submitted_at":"2026-03-15T02:21:05Z","abstract_excerpt":"Multimodal large language models (MLLMs) often suffer from perceptual impairments under extended reasoning modes, particularly in visual question answering (VQA) tasks. We identify attention dispersion as the underlying cause: during multi-step reasoning, the model's visual attention becomes scattered and drifts away from question-relevant regions, effectively \"losing focus\" on the visual input. To better understand this phenomenon, we analyze the attention maps of MLLMs and observe that reasoning prompts significantly reduce attention to regions critical for answering the question. We further"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2603.14184","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-03-15T02:21:05Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"767bf52d3b04d4b45680b6a9a6cf6db7bde58e08b77476a59d626db3a19363ae","abstract_canon_sha256":"d07f38ecb53f6de565b5fdfea5685e36fc42bfc429330f7b756bb48a04b43f8a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:05:17.757971Z","signature_b64":"yYJUJ0mQGuzv+oB7zTHaQ3SUaxG/smNQhR+uCrjPXDVmMB8p/aIbgRzsbT8SpvrWLaLw18rVOX4n9pkzcbpUDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f2cf232a738054222294dac09317e886372e61d304637af86f7764f89084f9a6","last_reissued_at":"2026-05-21T01:05:17.757273Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:05:17.757273Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Deeper Thought, Weaker Aim: Understanding and Mitigating Perceptual Impairment during Reasoning in Multimodal Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Jing Lei, Lu Hou, Ruiying Peng, Xiaohui Li, Xueyu Wu, Yuanzheng Ma","submitted_at":"2026-03-15T02:21:05Z","abstract_excerpt":"Multimodal large language models (MLLMs) often suffer from perceptual impairments under extended reasoning modes, particularly in visual question answering (VQA) tasks. We identify attention dispersion as the underlying cause: during multi-step reasoning, the model's visual attention becomes scattered and drifts away from question-relevant regions, effectively \"losing focus\" on the visual input. To better understand this phenomenon, we analyze the attention maps of MLLMs and observe that reasoning prompts significantly reduce attention to regions critical for answering the question. We further"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.14184","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2603.14184/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2603.14184","created_at":"2026-05-21T01:05:17.757379+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.14184v2","created_at":"2026-05-21T01:05:17.757379+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.14184","created_at":"2026-05-21T01:05:17.757379+00:00"},{"alias_kind":"pith_short_12","alias_value":"6LHSGKTTQBKC","created_at":"2026-05-21T01:05:17.757379+00:00"},{"alias_kind":"pith_short_16","alias_value":"6LHSGKTTQBKCEIUU","created_at":"2026-05-21T01:05:17.757379+00:00"},{"alias_kind":"pith_short_8","alias_value":"6LHSGKTT","created_at":"2026-05-21T01:05:17.757379+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6LHSGKTTQBKCEIUU3LAJGF7IQY","json":"https://pith.science/pith/6LHSGKTTQBKCEIUU3LAJGF7IQY.json","graph_json":"https://pith.science/api/pith-number/6LHSGKTTQBKCEIUU3LAJGF7IQY/graph.json","events_json":"https://pith.science/api/pith-number/6LHSGKTTQBKCEIUU3LAJGF7IQY/events.json","paper":"https://pith.science/paper/6LHSGKTT"},"agent_actions":{"view_html":"https://pith.science/pith/6LHSGKTTQBKCEIUU3LAJGF7IQY","download_json":"https://pith.science/pith/6LHSGKTTQBKCEIUU3LAJGF7IQY.json","view_paper":"https://pith.science/paper/6LHSGKTT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.14184&json=true","fetch_graph":"https://pith.science/api/pith-number/6LHSGKTTQBKCEIUU3LAJGF7IQY/graph.json","fetch_events":"https://pith.science/api/pith-number/6LHSGKTTQBKCEIUU3LAJGF7IQY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6LHSGKTTQBKCEIUU3LAJGF7IQY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6LHSGKTTQBKCEIUU3LAJGF7IQY/action/storage_attestation","attest_author":"https://pith.science/pith/6LHSGKTTQBKCEIUU3LAJGF7IQY/action/author_attestation","sign_citation":"https://pith.science/pith/6LHSGKTTQBKCEIUU3LAJGF7IQY/action/citation_signature","submit_replication":"https://pith.science/pith/6LHSGKTTQBKCEIUU3LAJGF7IQY/action/replication_record"}},"created_at":"2026-05-21T01:05:17.757379+00:00","updated_at":"2026-05-21T01:05:17.757379+00:00"}