{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:2REERM5RYU3SWPF3VPJ5PQHOQP","short_pith_number":"pith:2REERM5R","canonical_record":{"source":{"id":"2405.11165","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-05-18T03:49:37Z","cross_cats_sorted":[],"title_canon_sha256":"2d2361eafb62926a9700752e6d173d05e9e7450b278544c12e0561f9da06d565","abstract_canon_sha256":"e3d3c60edd299618e8f5b8d9735e6150ae0843247e302a9e343c204f82b9640c"},"schema_version":"1.0"},"canonical_sha256":"d44848b3b1c5372b3cbbabd3d7c0ee83ef4e80315dd48de6614b88fd78b06013","source":{"kind":"arxiv","id":"2405.11165","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.11165","created_at":"2026-07-05T08:24:24Z"},{"alias_kind":"arxiv_version","alias_value":"2405.11165v4","created_at":"2026-07-05T08:24:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.11165","created_at":"2026-07-05T08:24:24Z"},{"alias_kind":"pith_short_12","alias_value":"2REERM5RYU3S","created_at":"2026-07-05T08:24:24Z"},{"alias_kind":"pith_short_16","alias_value":"2REERM5RYU3SWPF3","created_at":"2026-07-05T08:24:24Z"},{"alias_kind":"pith_short_8","alias_value":"2REERM5R","created_at":"2026-07-05T08:24:24Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:2REERM5RYU3SWPF3VPJ5PQHOQP","target":"record","payload":{"canonical_record":{"source":{"id":"2405.11165","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-05-18T03:49:37Z","cross_cats_sorted":[],"title_canon_sha256":"2d2361eafb62926a9700752e6d173d05e9e7450b278544c12e0561f9da06d565","abstract_canon_sha256":"e3d3c60edd299618e8f5b8d9735e6150ae0843247e302a9e343c204f82b9640c"},"schema_version":"1.0"},"canonical_sha256":"d44848b3b1c5372b3cbbabd3d7c0ee83ef4e80315dd48de6614b88fd78b06013","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:24:24.897560Z","signature_b64":"Zg7L+bONHON9JBIqWNZ1IhKl13lxEgSlkgBKiG/2c0nTQhU/4WFYEWki3gSd0J72GceL3GV93FDECzkEXQCICQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d44848b3b1c5372b3cbbabd3d7c0ee83ef4e80315dd48de6614b88fd78b06013","last_reissued_at":"2026-07-05T08:24:24.897106Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:24:24.897106Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.11165","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:24:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"KqaXagPiS1FHjRqZwcQWufvNH1j23gfyfd6+UxPPfntzvaSdP0PCrZ9GlQ1gmQSAUtpWufDpkI2lZ+iKIEVXBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-22T15:38:02.676071Z"},"content_sha256":"2bfd313814a33f86139ba855eb9face9d033cd09cdd07ae65dac3a3a941a279f","schema_version":"1.0","event_id":"sha256:2bfd313814a33f86139ba855eb9face9d033cd09cdd07ae65dac3a3a941a279f"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:2REERM5RYU3SWPF3VPJ5PQHOQP","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Automated Multi-level Preference for MLLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Fanglong Liu, Haocheng Feng, Huanjin Yao, Jianbo Zhao, Jingdong Wang, Kang Rong, Mengxi Zhang, Wenhao Wu, Yifan Sun, Yu Lu, Yuxin Song","submitted_at":"2024-05-18T03:49:37Z","abstract_excerpt":"Current multimodal Large Language Models (MLLMs) suffer from ``hallucination'', occasionally generating responses that are not grounded in the input images. To tackle this challenge, one promising path is to utilize reinforcement learning from human feedback (RLHF), which steers MLLMs towards learning superior responses while avoiding inferior ones. We rethink the common practice of using binary preferences (i.e., superior, inferior), and find that adopting multi-level preferences (e.g., superior, medium, inferior) is better for two benefits: 1) It narrows the gap between adjacent levels, ther"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.11165","kind":"arxiv","version":4},"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/2405.11165/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":"2026-07-05T08:24:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"P0EfwskaRYCUHRrbnFY+zFfhL7AeuV5Ssn9QXPIH81Ki/jLT/mJV84tnf+aQsfwD/cbVOYBpw2jkVHGRhgVQDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-22T15:38:02.676452Z"},"content_sha256":"5e700c7738dd6dcefb4aa9c0524ac0ae55dce2baf7e09f8ae1d61e81e25674a6","schema_version":"1.0","event_id":"sha256:5e700c7738dd6dcefb4aa9c0524ac0ae55dce2baf7e09f8ae1d61e81e25674a6"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/2REERM5RYU3SWPF3VPJ5PQHOQP/bundle.json","state_url":"https://pith.science/pith/2REERM5RYU3SWPF3VPJ5PQHOQP/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/2REERM5RYU3SWPF3VPJ5PQHOQP/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":"2026-07-22T15:38:02Z","links":{"resolver":"https://pith.science/pith/2REERM5RYU3SWPF3VPJ5PQHOQP","bundle":"https://pith.science/pith/2REERM5RYU3SWPF3VPJ5PQHOQP/bundle.json","state":"https://pith.science/pith/2REERM5RYU3SWPF3VPJ5PQHOQP/state.json","well_known_bundle":"https://pith.science/.well-known/pith/2REERM5RYU3SWPF3VPJ5PQHOQP/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:2REERM5RYU3SWPF3VPJ5PQHOQP","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":"e3d3c60edd299618e8f5b8d9735e6150ae0843247e302a9e343c204f82b9640c","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-05-18T03:49:37Z","title_canon_sha256":"2d2361eafb62926a9700752e6d173d05e9e7450b278544c12e0561f9da06d565"},"schema_version":"1.0","source":{"id":"2405.11165","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.11165","created_at":"2026-07-05T08:24:24Z"},{"alias_kind":"arxiv_version","alias_value":"2405.11165v4","created_at":"2026-07-05T08:24:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.11165","created_at":"2026-07-05T08:24:24Z"},{"alias_kind":"pith_short_12","alias_value":"2REERM5RYU3S","created_at":"2026-07-05T08:24:24Z"},{"alias_kind":"pith_short_16","alias_value":"2REERM5RYU3SWPF3","created_at":"2026-07-05T08:24:24Z"},{"alias_kind":"pith_short_8","alias_value":"2REERM5R","created_at":"2026-07-05T08:24:24Z"}],"graph_snapshots":[{"event_id":"sha256:5e700c7738dd6dcefb4aa9c0524ac0ae55dce2baf7e09f8ae1d61e81e25674a6","target":"graph","created_at":"2026-07-05T08:24:24Z","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/2405.11165/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Current multimodal Large Language Models (MLLMs) suffer from ``hallucination'', occasionally generating responses that are not grounded in the input images. To tackle this challenge, one promising path is to utilize reinforcement learning from human feedback (RLHF), which steers MLLMs towards learning superior responses while avoiding inferior ones. We rethink the common practice of using binary preferences (i.e., superior, inferior), and find that adopting multi-level preferences (e.g., superior, medium, inferior) is better for two benefits: 1) It narrows the gap between adjacent levels, ther","authors_text":"Fanglong Liu, Haocheng Feng, Huanjin Yao, Jianbo Zhao, Jingdong Wang, Kang Rong, Mengxi Zhang, Wenhao Wu, Yifan Sun, Yu Lu, Yuxin Song","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-05-18T03:49:37Z","title":"Automated Multi-level Preference for MLLMs"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.11165","kind":"arxiv","version":4},"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":"sha256:2bfd313814a33f86139ba855eb9face9d033cd09cdd07ae65dac3a3a941a279f","target":"record","created_at":"2026-07-05T08:24:24Z","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":"e3d3c60edd299618e8f5b8d9735e6150ae0843247e302a9e343c204f82b9640c","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-05-18T03:49:37Z","title_canon_sha256":"2d2361eafb62926a9700752e6d173d05e9e7450b278544c12e0561f9da06d565"},"schema_version":"1.0","source":{"id":"2405.11165","kind":"arxiv","version":4}},"canonical_sha256":"d44848b3b1c5372b3cbbabd3d7c0ee83ef4e80315dd48de6614b88fd78b06013","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"d44848b3b1c5372b3cbbabd3d7c0ee83ef4e80315dd48de6614b88fd78b06013","first_computed_at":"2026-07-05T08:24:24.897106Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:24:24.897106Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Zg7L+bONHON9JBIqWNZ1IhKl13lxEgSlkgBKiG/2c0nTQhU/4WFYEWki3gSd0J72GceL3GV93FDECzkEXQCICQ==","signature_status":"signed_v1","signed_at":"2026-07-05T08:24:24.897560Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.11165","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:2bfd313814a33f86139ba855eb9face9d033cd09cdd07ae65dac3a3a941a279f","sha256:5e700c7738dd6dcefb4aa9c0524ac0ae55dce2baf7e09f8ae1d61e81e25674a6"],"state_sha256":"956381f2c5318036205c9601e636323f8730415ef7c4795668e10a8cb334ee0f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"BQRBdnf3upAP0m4ux0jzaX8KbPZzcSanR36eD4rI3xMKkyOsWl/STLNNZ7qLG/fBDfzN96g5t9ox94LkwUKVBg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-22T15:38:02.678515Z","bundle_sha256":"20ad11db512850ee9072ca5bbff6a10a74a76b7fc1f1286e08a1faba93ec313d"}}