{"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-2605-11591","title":"Logit-Attention Divergence: Mitigating Position Bias in Multi-Image Retrieval via Attention-Guided Calibration","arxiv_id":"2605.11591","date":"2026-05-12","proceeding":"ICML","authors":["Mingtao Xian","Yifeng Yang","Qinying Gu","Xinbing Wang","Nanyang Ye"],"abstract":"Multimodal Large Language Models (MLLMs) have shown strong performance in multi-image cross-modal retrieval, yet suffer from severe position bias, where predictions are dominated by input order rather than semantic relevance. Through empirical analysis, we identify a phenomenon termed Logit-Attention Divergence, in which output logits are heavily biased while internal attention maps remain well-aligned with relevant visual evidence. This observation reveals a fundamental limitation of existing logit-level calibration methods such as PriDe. Based on this insight, we propose a training-free, attention-guided debiasing framework that leverages intrinsic attention signals for instance-level correction at inference time, requiring only a minimal calibration set with negligible computational overhead. Experiments on MS-COCO-based benchmarks show that our method substantially improves permutation invariance and achieves state-of-the-art performance, enhancing accuracy by over 40\\% compared to baselines. Code is available at https://github.com/brightXian/LAD.","url_abs":"https://arxiv.org/abs/2605.11591","url_pdf":"https://arxiv.org/pdf/2605.11591","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":{"atlas_url":"https://app.syntology.ai/?focus=2605.11591","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2605.11591"}},"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":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/brightXian/LAD","reach":null}],"summary":{"ran":1,"ran_draft_wrong":3,"ran_honours":2,"unverified":6},"by_repo_kind":{"found_in_text":{"samples":12,"ran":6,"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":12,"samples":[{"code_sha256_prefix":"d58a9a0363d3436d","entry":"AttentionHook","repo":"brightXian/LAD","repo_kind":"found_in_text","path":"src/agd.py","file_url":"https://github.com/brightXian/LAD/blob/HEAD/src/agd.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"d58a9a0363d3436d"}},{"code_sha256_prefix":"43a61db67f8e8531","entry":"_calibration_messages","repo":"brightXian/LAD","repo_kind":"found_in_text","path":"src/agd.py","file_url":"https://github.com/brightXian/LAD/blob/HEAD/src/agd.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"43a61db67f8e8531"}},{"code_sha256_prefix":"80d2ee9043f65784","entry":"compute_bias_matrix","repo":"brightXian/LAD","repo_kind":"found_in_text","path":"src/agd.py","file_url":"https://github.com/brightXian/LAD/blob/HEAD/src/agd.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"80d2ee9043f65784"}},{"code_sha256_prefix":"37211fc5bf29b615","entry":"encode_hf_multimodal","repo":"brightXian/LAD","repo_kind":"found_in_text","path":"src/agd.py","file_url":"https://github.com/brightXian/LAD/blob/HEAD/src/agd.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"37211fc5bf29b615"}},{"code_sha256_prefix":"498a44017a02f6b8","entry":"find_image_token_ranges","repo":"brightXian/LAD","repo_kind":"found_in_text","path":"src/agd.py","file_url":"https://github.com/brightXian/LAD/blob/HEAD/src/agd.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"498a44017a02f6b8"}},{"code_sha256_prefix":"f53971fbc29934cc","entry":"process_images","repo":"brightXian/LAD","repo_kind":"found_in_text","path":"src/agd.py","file_url":"https://github.com/brightXian/LAD/blob/HEAD/src/agd.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"f53971fbc29934cc"}},{"code_sha256_prefix":"6014728bf8b76c40","entry":"aggregate_layer_scores","repo":"brightXian/LAD","repo_kind":"found_in_text","path":"src/agd.py","file_url":"https://github.com/brightXian/LAD/blob/HEAD/src/agd.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"6014728bf8b76c40"}},{"code_sha256_prefix":"9a76a397136006ca","entry":"collect_logits_stats","repo":"brightXian/LAD","repo_kind":"found_in_text","path":"src/agd.py","file_url":"https://github.com/brightXian/LAD/blob/HEAD/src/agd.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"9a76a397136006ca"}},{"code_sha256_prefix":"c8ba677830a98248","entry":"compute_joint_log_probs","repo":"brightXian/LAD","repo_kind":"found_in_text","path":"src/agd.py","file_url":"https://github.com/brightXian/LAD/blob/HEAD/src/agd.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"c8ba677830a98248"}},{"code_sha256_prefix":"da22f55acfa736b9","entry":"compute_restricted_log_probs","repo":"brightXian/LAD","repo_kind":"found_in_text","path":"src/agd.py","file_url":"https://github.com/brightXian/LAD/blob/HEAD/src/agd.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"da22f55acfa736b9"}},{"code_sha256_prefix":"784c5edfc3477e70","entry":"get_joint_log_probs","repo":"brightXian/LAD","repo_kind":"found_in_text","path":"src/agd.py","file_url":"https://github.com/brightXian/LAD/blob/HEAD/src/agd.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"784c5edfc3477e70"}},{"code_sha256_prefix":"862e0915605ca744","entry":"run_agd_calibration","repo":"brightXian/LAD","repo_kind":"found_in_text","path":"src/agd.py","file_url":"https://github.com/brightXian/LAD/blob/HEAD/src/agd.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"862e0915605ca744"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.CV","source":"arxiv_2026.jsonl"},"syntology_extracted_results":null}