{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:B3DSA7YXYW3DEXDHVCG3BSFCVS","short_pith_number":"pith:B3DSA7YX","canonical_record":{"source":{"id":"2608.11758","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-08-12T07:55:15Z","cross_cats_sorted":[],"title_canon_sha256":"19e82ea3ea288d4328fb454d52186a55860d5a11719c9c5da8764eebb58ad53a","abstract_canon_sha256":"b5ba6a0f20ce64e62fb02c3e274c7cf599b6d039ed29c656947513a573cfc7a5"},"schema_version":"1.0"},"canonical_sha256":"0ec7207f17c5b6325c67a88db0c8a2acba7c9e388e6494ad9fce9dceb87890c7","source":{"kind":"arxiv","id":"2608.11758","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2608.11758","created_at":"2026-08-13T01:28:06Z"},{"alias_kind":"arxiv_version","alias_value":"2608.11758v1","created_at":"2026-08-13T01:28:06Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.11758","created_at":"2026-08-13T01:28:06Z"},{"alias_kind":"pith_short_12","alias_value":"B3DSA7YXYW3D","created_at":"2026-08-13T01:28:06Z"},{"alias_kind":"pith_short_16","alias_value":"B3DSA7YXYW3DEXDH","created_at":"2026-08-13T01:28:06Z"},{"alias_kind":"pith_short_8","alias_value":"B3DSA7YX","created_at":"2026-08-13T01:28:06Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:B3DSA7YXYW3DEXDHVCG3BSFCVS","target":"record","payload":{"canonical_record":{"source":{"id":"2608.11758","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-08-12T07:55:15Z","cross_cats_sorted":[],"title_canon_sha256":"19e82ea3ea288d4328fb454d52186a55860d5a11719c9c5da8764eebb58ad53a","abstract_canon_sha256":"b5ba6a0f20ce64e62fb02c3e274c7cf599b6d039ed29c656947513a573cfc7a5"},"schema_version":"1.0"},"canonical_sha256":"0ec7207f17c5b6325c67a88db0c8a2acba7c9e388e6494ad9fce9dceb87890c7","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-13T01:28:06.669058Z","signature_b64":"a8ouAqJXc32EJ2EmvV100TM1ELFa3YHyafLuokQqeDPuDwZhw2SBavbvcNSxUja3/sPegMMTObMhWnq2pMZ1Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0ec7207f17c5b6325c67a88db0c8a2acba7c9e388e6494ad9fce9dceb87890c7","last_reissued_at":"2026-08-13T01:28:06.666166Z","signature_status":"signed_v1","first_computed_at":"2026-08-13T01:28:06.666166Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2608.11758","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":"2026-08-13T01:28:06Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"PCeqE7jcgKezJ3peX8X8Ms1wFgEe7zYqveSYcQSCKWiBBOnTEuKTqho/4cBkRvwVjYtRLz9lfK4SiVdTgkDuDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-18T08:38:33.178309Z"},"content_sha256":"8126fdb4cf4a64deb7ef41aa908c5b2265f10bbebac4a088bfebdbcc16b8d3fa","schema_version":"1.0","event_id":"sha256:8126fdb4cf4a64deb7ef41aa908c5b2265f10bbebac4a088bfebdbcc16b8d3fa"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:B3DSA7YXYW3DEXDHVCG3BSFCVS","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"AWARe: Mitigating Catastrophic Forgetting via Activation-Weighted Adaptive REtention","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Guoming Wang, Jinfan Lv, Juncheng Liao, Jupeng Zheng, Ling Xiao, Siliang Tang","submitted_at":"2026-08-12T07:55:15Z","abstract_excerpt":"Multimodal Large Language Models (MLLMs) exhibit strong generalization and reasoning abilities due to large-scale multimodal pre-training. However, fine-tuning these models on downstream tasks often leads to catastrophic forgetting, where newly learned task-specific knowledge degrades previously acquired capabilities. This issue arises because gradient updates for new tasks overwrite parameters critical to prior knowledge, limiting the practical deployment of MLLMs. To address this challenge, we propose Activation-Weighted Adaptive REtention (AWARe), a fine-tuning method that mitigates catastr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.11758","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/2608.11758/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":"2026-08-13T01:28:06Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"bw1LWMl835su7MZ0GyzCN1nzYRrkuqaeh0PADVWIC3DDyZ9lNSO5GaeBJ2xaarsHqKqCIDcxI8LHpWXzrfYABA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-18T08:38:33.182842Z"},"content_sha256":"909c7d17d6e84bcb168fa748b97342eab61c8ef80659e4c2071711298c24e952","schema_version":"1.0","event_id":"sha256:909c7d17d6e84bcb168fa748b97342eab61c8ef80659e4c2071711298c24e952"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/B3DSA7YXYW3DEXDHVCG3BSFCVS/bundle.json","state_url":"https://pith.science/pith/B3DSA7YXYW3DEXDHVCG3BSFCVS/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/B3DSA7YXYW3DEXDHVCG3BSFCVS/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":"2026-08-18T08:38:33Z","links":{"resolver":"https://pith.science/pith/B3DSA7YXYW3DEXDHVCG3BSFCVS","bundle":"https://pith.science/pith/B3DSA7YXYW3DEXDHVCG3BSFCVS/bundle.json","state":"https://pith.science/pith/B3DSA7YXYW3DEXDHVCG3BSFCVS/state.json","well_known_bundle":"https://pith.science/.well-known/pith/B3DSA7YXYW3DEXDHVCG3BSFCVS/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:B3DSA7YXYW3DEXDHVCG3BSFCVS","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":"b5ba6a0f20ce64e62fb02c3e274c7cf599b6d039ed29c656947513a573cfc7a5","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-08-12T07:55:15Z","title_canon_sha256":"19e82ea3ea288d4328fb454d52186a55860d5a11719c9c5da8764eebb58ad53a"},"schema_version":"1.0","source":{"id":"2608.11758","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2608.11758","created_at":"2026-08-13T01:28:06Z"},{"alias_kind":"arxiv_version","alias_value":"2608.11758v1","created_at":"2026-08-13T01:28:06Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.11758","created_at":"2026-08-13T01:28:06Z"},{"alias_kind":"pith_short_12","alias_value":"B3DSA7YXYW3D","created_at":"2026-08-13T01:28:06Z"},{"alias_kind":"pith_short_16","alias_value":"B3DSA7YXYW3DEXDH","created_at":"2026-08-13T01:28:06Z"},{"alias_kind":"pith_short_8","alias_value":"B3DSA7YX","created_at":"2026-08-13T01:28:06Z"}],"graph_snapshots":[{"event_id":"sha256:909c7d17d6e84bcb168fa748b97342eab61c8ef80659e4c2071711298c24e952","target":"graph","created_at":"2026-08-13T01:28:06Z","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/2608.11758/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Multimodal Large Language Models (MLLMs) exhibit strong generalization and reasoning abilities due to large-scale multimodal pre-training. However, fine-tuning these models on downstream tasks often leads to catastrophic forgetting, where newly learned task-specific knowledge degrades previously acquired capabilities. This issue arises because gradient updates for new tasks overwrite parameters critical to prior knowledge, limiting the practical deployment of MLLMs. To address this challenge, we propose Activation-Weighted Adaptive REtention (AWARe), a fine-tuning method that mitigates catastr","authors_text":"Guoming Wang, Jinfan Lv, Juncheng Liao, Jupeng Zheng, Ling Xiao, Siliang Tang","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-08-12T07:55:15Z","title":"AWARe: Mitigating Catastrophic Forgetting via Activation-Weighted Adaptive REtention"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.11758","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":"sha256:8126fdb4cf4a64deb7ef41aa908c5b2265f10bbebac4a088bfebdbcc16b8d3fa","target":"record","created_at":"2026-08-13T01:28:06Z","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":"b5ba6a0f20ce64e62fb02c3e274c7cf599b6d039ed29c656947513a573cfc7a5","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-08-12T07:55:15Z","title_canon_sha256":"19e82ea3ea288d4328fb454d52186a55860d5a11719c9c5da8764eebb58ad53a"},"schema_version":"1.0","source":{"id":"2608.11758","kind":"arxiv","version":1}},"canonical_sha256":"0ec7207f17c5b6325c67a88db0c8a2acba7c9e388e6494ad9fce9dceb87890c7","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0ec7207f17c5b6325c67a88db0c8a2acba7c9e388e6494ad9fce9dceb87890c7","first_computed_at":"2026-08-13T01:28:06.666166Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-08-13T01:28:06.666166Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"a8ouAqJXc32EJ2EmvV100TM1ELFa3YHyafLuokQqeDPuDwZhw2SBavbvcNSxUja3/sPegMMTObMhWnq2pMZ1Aw==","signature_status":"signed_v1","signed_at":"2026-08-13T01:28:06.669058Z","signed_message":"canonical_sha256_bytes"},"source_id":"2608.11758","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:8126fdb4cf4a64deb7ef41aa908c5b2265f10bbebac4a088bfebdbcc16b8d3fa","sha256:909c7d17d6e84bcb168fa748b97342eab61c8ef80659e4c2071711298c24e952"],"state_sha256":"6d0e51b82cb6ef0c7663456f42d7d0b706129baf7883db1ee6e4dc1bb31807dc"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"mgbx2yoKM9nFl4q89jYufF/bf/87jtQhUE6RLKhZmidUpZ6PPauNz8wkKBVMAOlVJ5FMH+Yhym468VrKonhMAQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-18T08:38:33.193212Z","bundle_sha256":"1d24903cf5aeb93ef4e8eddaf508a7cccf764a4e6f38b36c0d01f5d9cb586fc9"}}