{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:X6OAUBAU6GZU63RGUWUMHEWBTK","short_pith_number":"pith:X6OAUBAU","canonical_record":{"source":{"id":"2605.25036","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-24T12:23:13Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8af0d2f26bb43c0c6a2483a40719e0ff4434b1f9eb22390dce23aa4b4c86b785","abstract_canon_sha256":"d436ad0eea11765a1ff900427d1915cdae5dd0425b185097e6ee2ac42829394b"},"schema_version":"1.0"},"canonical_sha256":"bf9c0a0414f1b34f6e26a5a8c392c19abc88682c434e988b7d7a89d90edf9e25","source":{"kind":"arxiv","id":"2605.25036","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.25036","created_at":"2026-05-26T02:03:36Z"},{"alias_kind":"arxiv_version","alias_value":"2605.25036v1","created_at":"2026-05-26T02:03:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.25036","created_at":"2026-05-26T02:03:36Z"},{"alias_kind":"pith_short_12","alias_value":"X6OAUBAU6GZU","created_at":"2026-05-26T02:03:36Z"},{"alias_kind":"pith_short_16","alias_value":"X6OAUBAU6GZU63RG","created_at":"2026-05-26T02:03:36Z"},{"alias_kind":"pith_short_8","alias_value":"X6OAUBAU","created_at":"2026-05-26T02:03:36Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:X6OAUBAU6GZU63RGUWUMHEWBTK","target":"record","payload":{"canonical_record":{"source":{"id":"2605.25036","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-24T12:23:13Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8af0d2f26bb43c0c6a2483a40719e0ff4434b1f9eb22390dce23aa4b4c86b785","abstract_canon_sha256":"d436ad0eea11765a1ff900427d1915cdae5dd0425b185097e6ee2ac42829394b"},"schema_version":"1.0"},"canonical_sha256":"bf9c0a0414f1b34f6e26a5a8c392c19abc88682c434e988b7d7a89d90edf9e25","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T02:03:36.658941Z","signature_b64":"Eqaa6wzV06QxWDCMRzQpNtUeeuUCY9Bn6PisMlsOhbBm2Rrg3hH8wzLM2Z6XmsPsC6HkivHnl7gbEUILLNNqCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bf9c0a0414f1b34f6e26a5a8c392c19abc88682c434e988b7d7a89d90edf9e25","last_reissued_at":"2026-05-26T02:03:36.658095Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T02:03:36.658095Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.25036","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-26T02:03:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8XWg4S0VIBoVwtlnt5PJ2sN2JEUYMtqTEcPKalRJ3XTf9L7hvh0h6OA2hol/TGHjgfYf3dMEKbix6xpZwcExAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T08:07:37.620309Z"},"content_sha256":"b064bde2886962b301cedc01d81771ce545fb37b7d97ee6bccd04737a661c089","schema_version":"1.0","event_id":"sha256:b064bde2886962b301cedc01d81771ce545fb37b7d97ee6bccd04737a661c089"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:X6OAUBAU6GZU63RGUWUMHEWBTK","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Language Bias in LVLMs: From In-Depth Analysis to Simple and Effective Mitigation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Jing Li, Yangneng Chen","submitted_at":"2026-05-24T12:23:13Z","abstract_excerpt":"Large Vision-Language Models (LVLMs) extend large language models with visual understanding, but remain vulnerable to hallucination, where outputs are fluent yet inconsistent with images. Recent studies link this issue to language bias-the tendency of LVLMs to over-rely on text while neglecting visual inputs. Yet most analyses remain empirical without uncovering its underlying cause. In this paper, we provide a systematic study of language bias and identify its root in modality misalignment during training. Our analysis shows that both Visual Instruction Tuning (VIT) and Direct Preference Opti"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.25036","kind":"arxiv","version":1},"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/2605.25036/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-26T02:03:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"co+jIRojCbtS08zogjUrHhtoz2caKD81jlStaSYEpibuBJSHvb2llZGiP8nAGFrGo0UyLsfvPAY+TpDffdS3Bg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T08:07:37.620823Z"},"content_sha256":"b028f0346d5fd8b714b262216c4761acdc6277ca06b134e92d3d99ec01f835d7","schema_version":"1.0","event_id":"sha256:b028f0346d5fd8b714b262216c4761acdc6277ca06b134e92d3d99ec01f835d7"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/X6OAUBAU6GZU63RGUWUMHEWBTK/bundle.json","state_url":"https://pith.science/pith/X6OAUBAU6GZU63RGUWUMHEWBTK/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/X6OAUBAU6GZU63RGUWUMHEWBTK/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-07T08:07:37Z","links":{"resolver":"https://pith.science/pith/X6OAUBAU6GZU63RGUWUMHEWBTK","bundle":"https://pith.science/pith/X6OAUBAU6GZU63RGUWUMHEWBTK/bundle.json","state":"https://pith.science/pith/X6OAUBAU6GZU63RGUWUMHEWBTK/state.json","well_known_bundle":"https://pith.science/.well-known/pith/X6OAUBAU6GZU63RGUWUMHEWBTK/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:X6OAUBAU6GZU63RGUWUMHEWBTK","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"d436ad0eea11765a1ff900427d1915cdae5dd0425b185097e6ee2ac42829394b","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-24T12:23:13Z","title_canon_sha256":"8af0d2f26bb43c0c6a2483a40719e0ff4434b1f9eb22390dce23aa4b4c86b785"},"schema_version":"1.0","source":{"id":"2605.25036","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.25036","created_at":"2026-05-26T02:03:36Z"},{"alias_kind":"arxiv_version","alias_value":"2605.25036v1","created_at":"2026-05-26T02:03:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.25036","created_at":"2026-05-26T02:03:36Z"},{"alias_kind":"pith_short_12","alias_value":"X6OAUBAU6GZU","created_at":"2026-05-26T02:03:36Z"},{"alias_kind":"pith_short_16","alias_value":"X6OAUBAU6GZU63RG","created_at":"2026-05-26T02:03:36Z"},{"alias_kind":"pith_short_8","alias_value":"X6OAUBAU","created_at":"2026-05-26T02:03:36Z"}],"graph_snapshots":[{"event_id":"sha256:b028f0346d5fd8b714b262216c4761acdc6277ca06b134e92d3d99ec01f835d7","target":"graph","created_at":"2026-05-26T02:03:36Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2605.25036/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Vision-Language Models (LVLMs) extend large language models with visual understanding, but remain vulnerable to hallucination, where outputs are fluent yet inconsistent with images. Recent studies link this issue to language bias-the tendency of LVLMs to over-rely on text while neglecting visual inputs. Yet most analyses remain empirical without uncovering its underlying cause. In this paper, we provide a systematic study of language bias and identify its root in modality misalignment during training. Our analysis shows that both Visual Instruction Tuning (VIT) and Direct Preference Opti","authors_text":"Jing Li, Yangneng Chen","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-24T12:23:13Z","title":"Language Bias in LVLMs: From In-Depth Analysis to Simple and Effective Mitigation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.25036","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:b064bde2886962b301cedc01d81771ce545fb37b7d97ee6bccd04737a661c089","target":"record","created_at":"2026-05-26T02:03:36Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"d436ad0eea11765a1ff900427d1915cdae5dd0425b185097e6ee2ac42829394b","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-24T12:23:13Z","title_canon_sha256":"8af0d2f26bb43c0c6a2483a40719e0ff4434b1f9eb22390dce23aa4b4c86b785"},"schema_version":"1.0","source":{"id":"2605.25036","kind":"arxiv","version":1}},"canonical_sha256":"bf9c0a0414f1b34f6e26a5a8c392c19abc88682c434e988b7d7a89d90edf9e25","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"bf9c0a0414f1b34f6e26a5a8c392c19abc88682c434e988b7d7a89d90edf9e25","first_computed_at":"2026-05-26T02:03:36.658095Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-26T02:03:36.658095Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Eqaa6wzV06QxWDCMRzQpNtUeeuUCY9Bn6PisMlsOhbBm2Rrg3hH8wzLM2Z6XmsPsC6HkivHnl7gbEUILLNNqCg==","signature_status":"signed_v1","signed_at":"2026-05-26T02:03:36.658941Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.25036","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b064bde2886962b301cedc01d81771ce545fb37b7d97ee6bccd04737a661c089","sha256:b028f0346d5fd8b714b262216c4761acdc6277ca06b134e92d3d99ec01f835d7"],"state_sha256":"df6000e6b457830c205e70c1ac9e688352a8d08402d46c0f1f89b2913413bc49"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"TjwsBA32/k9D7EbgmOXz+kt2/iQwNLT3XJuX+LwRefL0GaGT0AdWWyHcDWZDVxbWrf0tnmdSAO0r92mduBQqAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T08:07:37.626332Z","bundle_sha256":"37dd3d5369d8a0e7b2816227d775444ac8acd7b88ddade6a94521adcc77e6609"}}