{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:CGZIDMV6YMQFYIINM25LY34XUC","short_pith_number":"pith:CGZIDMV6","canonical_record":{"source":{"id":"2312.17345","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2023-12-28T20:26:03Z","cross_cats_sorted":[],"title_canon_sha256":"85f4d2575a8045b6d79641d839523eaa325c9b260d3a58e24a4b006f4db1810f","abstract_canon_sha256":"82ce75bd3dc7949d2f6f6464a998402fa4ed851931dc9a3652a41b94ce286d08"},"schema_version":"1.0"},"canonical_sha256":"11b281b2bec3205c210d66babc6f97a086c6a9436e38dc836059e4eaaa525fbe","source":{"kind":"arxiv","id":"2312.17345","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2312.17345","created_at":"2026-07-05T10:01:11Z"},{"alias_kind":"arxiv_version","alias_value":"2312.17345v2","created_at":"2026-07-05T10:01:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.17345","created_at":"2026-07-05T10:01:11Z"},{"alias_kind":"pith_short_12","alias_value":"CGZIDMV6YMQF","created_at":"2026-07-05T10:01:11Z"},{"alias_kind":"pith_short_16","alias_value":"CGZIDMV6YMQFYIIN","created_at":"2026-07-05T10:01:11Z"},{"alias_kind":"pith_short_8","alias_value":"CGZIDMV6","created_at":"2026-07-05T10:01:11Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:CGZIDMV6YMQFYIINM25LY34XUC","target":"record","payload":{"canonical_record":{"source":{"id":"2312.17345","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2023-12-28T20:26:03Z","cross_cats_sorted":[],"title_canon_sha256":"85f4d2575a8045b6d79641d839523eaa325c9b260d3a58e24a4b006f4db1810f","abstract_canon_sha256":"82ce75bd3dc7949d2f6f6464a998402fa4ed851931dc9a3652a41b94ce286d08"},"schema_version":"1.0"},"canonical_sha256":"11b281b2bec3205c210d66babc6f97a086c6a9436e38dc836059e4eaaa525fbe","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:01:11.266351Z","signature_b64":"QXmjMYMqoRjaPF1Y1Xd4L4I2aTyMMjRjCFveYheBfkspdozYBAU/gmwbkuHO7UAkPu/HUHLee1EAT+/QJfSyBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"11b281b2bec3205c210d66babc6f97a086c6a9436e38dc836059e4eaaa525fbe","last_reissued_at":"2026-07-05T10:01:11.265879Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:01:11.265879Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2312.17345","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:01:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"iNGZjahPYQOO++CWsk/Fjthr4C3TNMrCIE3kfZlaZYiYOB15cmiIqfzfj2BuP62UxX1GGIx9sOwcKHdfIlXODw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-23T17:57:53.350065Z"},"content_sha256":"efe600e41bf3677563b4534f3cbfdad550e3b3d0354820e1b0f44e16a78b542c","schema_version":"1.0","event_id":"sha256:efe600e41bf3677563b4534f3cbfdad550e3b3d0354820e1b0f44e16a78b542c"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:CGZIDMV6YMQFYIINM25LY34XUC","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"3VL: Using Trees to Improve Vision-Language Models' Interpretability","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Leonid Karlinsky, Nir Yellinek, Raja Giryes","submitted_at":"2023-12-28T20:26:03Z","abstract_excerpt":"Vision-Language models (VLMs) have proven to be effective at aligning image and text representations, producing superior zero-shot results when transferred to many downstream tasks. However, these representations suffer from some key shortcomings in understanding Compositional Language Concepts (CLC), such as recognizing objects' attributes, states, and relations between different objects. Moreover, VLMs typically have poor interpretability, making it challenging to debug and mitigate compositional-understanding failures. In this work, we introduce the architecture and training technique of Tr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.17345","kind":"arxiv","version":2},"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/2312.17345/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":"2026-07-05T10:01:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/AenV/8mOQTjBOPWExesWENqiRNYsNfRCN0PkncP0S4wi5w9m1gHDa33VfrSWopZgrfYQJQhSF47L0FuBfziBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-23T17:57:53.350589Z"},"content_sha256":"d00222ef8222502d804b8bee0c8d81e038b9af8e895ad5c099886ea171edb423","schema_version":"1.0","event_id":"sha256:d00222ef8222502d804b8bee0c8d81e038b9af8e895ad5c099886ea171edb423"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/CGZIDMV6YMQFYIINM25LY34XUC/bundle.json","state_url":"https://pith.science/pith/CGZIDMV6YMQFYIINM25LY34XUC/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/CGZIDMV6YMQFYIINM25LY34XUC/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":"2026-08-23T17:57:53Z","links":{"resolver":"https://pith.science/pith/CGZIDMV6YMQFYIINM25LY34XUC","bundle":"https://pith.science/pith/CGZIDMV6YMQFYIINM25LY34XUC/bundle.json","state":"https://pith.science/pith/CGZIDMV6YMQFYIINM25LY34XUC/state.json","well_known_bundle":"https://pith.science/.well-known/pith/CGZIDMV6YMQFYIINM25LY34XUC/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:CGZIDMV6YMQFYIINM25LY34XUC","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":"82ce75bd3dc7949d2f6f6464a998402fa4ed851931dc9a3652a41b94ce286d08","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2023-12-28T20:26:03Z","title_canon_sha256":"85f4d2575a8045b6d79641d839523eaa325c9b260d3a58e24a4b006f4db1810f"},"schema_version":"1.0","source":{"id":"2312.17345","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2312.17345","created_at":"2026-07-05T10:01:11Z"},{"alias_kind":"arxiv_version","alias_value":"2312.17345v2","created_at":"2026-07-05T10:01:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.17345","created_at":"2026-07-05T10:01:11Z"},{"alias_kind":"pith_short_12","alias_value":"CGZIDMV6YMQF","created_at":"2026-07-05T10:01:11Z"},{"alias_kind":"pith_short_16","alias_value":"CGZIDMV6YMQFYIIN","created_at":"2026-07-05T10:01:11Z"},{"alias_kind":"pith_short_8","alias_value":"CGZIDMV6","created_at":"2026-07-05T10:01:11Z"}],"graph_snapshots":[{"event_id":"sha256:d00222ef8222502d804b8bee0c8d81e038b9af8e895ad5c099886ea171edb423","target":"graph","created_at":"2026-07-05T10:01:11Z","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/2312.17345/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Vision-Language models (VLMs) have proven to be effective at aligning image and text representations, producing superior zero-shot results when transferred to many downstream tasks. However, these representations suffer from some key shortcomings in understanding Compositional Language Concepts (CLC), such as recognizing objects' attributes, states, and relations between different objects. Moreover, VLMs typically have poor interpretability, making it challenging to debug and mitigate compositional-understanding failures. In this work, we introduce the architecture and training technique of Tr","authors_text":"Leonid Karlinsky, Nir Yellinek, Raja Giryes","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2023-12-28T20:26:03Z","title":"3VL: Using Trees to Improve Vision-Language Models' Interpretability"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.17345","kind":"arxiv","version":2},"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:efe600e41bf3677563b4534f3cbfdad550e3b3d0354820e1b0f44e16a78b542c","target":"record","created_at":"2026-07-05T10:01:11Z","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":"82ce75bd3dc7949d2f6f6464a998402fa4ed851931dc9a3652a41b94ce286d08","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2023-12-28T20:26:03Z","title_canon_sha256":"85f4d2575a8045b6d79641d839523eaa325c9b260d3a58e24a4b006f4db1810f"},"schema_version":"1.0","source":{"id":"2312.17345","kind":"arxiv","version":2}},"canonical_sha256":"11b281b2bec3205c210d66babc6f97a086c6a9436e38dc836059e4eaaa525fbe","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"11b281b2bec3205c210d66babc6f97a086c6a9436e38dc836059e4eaaa525fbe","first_computed_at":"2026-07-05T10:01:11.265879Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:01:11.265879Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"QXmjMYMqoRjaPF1Y1Xd4L4I2aTyMMjRjCFveYheBfkspdozYBAU/gmwbkuHO7UAkPu/HUHLee1EAT+/QJfSyBg==","signature_status":"signed_v1","signed_at":"2026-07-05T10:01:11.266351Z","signed_message":"canonical_sha256_bytes"},"source_id":"2312.17345","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:efe600e41bf3677563b4534f3cbfdad550e3b3d0354820e1b0f44e16a78b542c","sha256:d00222ef8222502d804b8bee0c8d81e038b9af8e895ad5c099886ea171edb423"],"state_sha256":"1c6e5a23f2f092e00a8acd937064029c3055e4ed40f8b553f0b8de9ceedb6bf5"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"cZLffOeMb1q1X/y7QEHwWEP6csLHEfRm6rwDiCzfgwO+W0BF9BtkBLNc6hAywmmt+U4FktP61hLi1/a9XHkYCw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-23T17:57:53.356926Z","bundle_sha256":"f7e4104aeaffbf9032535b981843d54da180b4be49d1ff1a99affd432c92c6d9"}}