{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:S7VVMMSIJIWELJMQ6JJ7ULDKS2","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":"0150563e79b100616c8eb00b2ac02a9864c67d53ec79342ccdc0543cf6860816","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-05T17:59:40Z","title_canon_sha256":"8ff0d71f5f0bb14daae58507aa629c86e7f5370866bdca8bf7a85725e363b1dd"},"schema_version":"1.0","source":{"id":"2406.03496","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.03496","created_at":"2026-07-05T08:27:58Z"},{"alias_kind":"arxiv_version","alias_value":"2406.03496v1","created_at":"2026-07-05T08:27:58Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.03496","created_at":"2026-07-05T08:27:58Z"},{"alias_kind":"pith_short_12","alias_value":"S7VVMMSIJIWE","created_at":"2026-07-05T08:27:58Z"},{"alias_kind":"pith_short_16","alias_value":"S7VVMMSIJIWELJMQ","created_at":"2026-07-05T08:27:58Z"},{"alias_kind":"pith_short_8","alias_value":"S7VVMMSI","created_at":"2026-07-05T08:27:58Z"}],"graph_snapshots":[{"event_id":"sha256:cfbac50a8852e58cc7fc3d823a7bb1ab178f1a6843185dff3ed1e78b852f1ce6","target":"graph","created_at":"2026-07-05T08:27:58Z","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/2406.03496/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Multimodal large language models (MLLMs), initiated with a trained LLM, first align images with text and then fine-tune on multimodal mixed inputs. However, the MLLM catastrophically forgets the text-only instructions, which do not include images and can be addressed within the initial LLM. In this paper, we present Wings, a novel MLLM that excels in both text-only dialogues and multimodal comprehension. Analyzing MLLM attention in multimodal instructions reveals that text-only forgetting is related to the attention shifts from pre-image to post-image text. From that, we construct extra module","authors_text":"De-Chuan Zhan, Han-Jia Ye, Kaifu Zhang, Qing-Guo Chen, Shiyin Lu, Weihua Luo, Yang Li, Yanqing Ma, Yi-Kai Zhang, Zhao Xu","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-05T17:59:40Z","title":"Wings: Learning Multimodal LLMs without Text-only Forgetting"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.03496","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:af8a5ca7bb4229d848d185cf9d3cdf0e396ba24ebc5c5b832f6df160bb05d5c1","target":"record","created_at":"2026-07-05T08:27:58Z","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":"0150563e79b100616c8eb00b2ac02a9864c67d53ec79342ccdc0543cf6860816","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-05T17:59:40Z","title_canon_sha256":"8ff0d71f5f0bb14daae58507aa629c86e7f5370866bdca8bf7a85725e363b1dd"},"schema_version":"1.0","source":{"id":"2406.03496","kind":"arxiv","version":1}},"canonical_sha256":"97eb5632484a2c45a590f253fa2c6a96836ddcb499a0d45486327a6bf1c14baf","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"97eb5632484a2c45a590f253fa2c6a96836ddcb499a0d45486327a6bf1c14baf","first_computed_at":"2026-07-05T08:27:58.472724Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:27:58.472724Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Hp2mcW1qiVhwclq4GxR5RUg+CM1w4f4OStcnnN0x+o33+oHw6NZnu3psLDguxdFx+8ZCflEifou7PwReZMiSCw==","signature_status":"signed_v1","signed_at":"2026-07-05T08:27:58.473208Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.03496","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:af8a5ca7bb4229d848d185cf9d3cdf0e396ba24ebc5c5b832f6df160bb05d5c1","sha256:cfbac50a8852e58cc7fc3d823a7bb1ab178f1a6843185dff3ed1e78b852f1ce6"],"state_sha256":"39b347a32e3311093621b77fa6308676599a445df1140534fffb0936b265e61f"}