{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:X43HB4EKVQEQSVRBMYIO4D5W7M","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":"49e8b665c03f258b823eb8adbac78fe2dfc2dcd1d9986a6de9774aab1f8e5d5f","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-06-05T16:54:40Z","title_canon_sha256":"d26a6241e95135087f703797b43419f63fb6a86c118efd5bd4b48c51f04cc1a6"},"schema_version":"1.0","source":{"id":"2606.07451","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.07451","created_at":"2026-06-08T01:05:27Z"},{"alias_kind":"arxiv_version","alias_value":"2606.07451v1","created_at":"2026-06-08T01:05:27Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.07451","created_at":"2026-06-08T01:05:27Z"},{"alias_kind":"pith_short_12","alias_value":"X43HB4EKVQEQ","created_at":"2026-06-08T01:05:27Z"},{"alias_kind":"pith_short_16","alias_value":"X43HB4EKVQEQSVRB","created_at":"2026-06-08T01:05:27Z"},{"alias_kind":"pith_short_8","alias_value":"X43HB4EK","created_at":"2026-06-08T01:05:27Z"}],"graph_snapshots":[{"event_id":"sha256:b9af860e7dfde09a14b39ad74fe9d78897b861fd826a7dec30f821311ea3ae5c","target":"graph","created_at":"2026-06-08T01:05:27Z","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/2606.07451/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Vision-language models such as CLIP are highly useful for diverse tasks due to their shared image-text embedding space. Despite this, the image and text embeddings are often poorly aligned, affecting downstream performance. Recent work has shown that this can be attributed to an information imbalance: images contain more information than their captions describe. In this work, we propose TEVI, a framework that uses captions as a signal for what to retain from image embeddings. Specifically, we use sparse autoencoders to disentangle image embeddings and train a masking module to selectively reco","authors_text":"Alexander Koller, Bernt Schiele, Jiahao Xie, Sukrut Rao, Sweta Mahajan","cross_cats":["cs.AI","cs.CL","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-06-05T16:54:40Z","title":"TEVI: Text-Conditioned Editing of Visual Representations via Sparse Autoencoders for Improved Vision-Language Alignment"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.07451","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":"sha256:7f0de5be742d321b3a45d2111a8dbd823123f87985bbf30520ebaed09435718b","target":"record","created_at":"2026-06-08T01:05:27Z","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":"49e8b665c03f258b823eb8adbac78fe2dfc2dcd1d9986a6de9774aab1f8e5d5f","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-06-05T16:54:40Z","title_canon_sha256":"d26a6241e95135087f703797b43419f63fb6a86c118efd5bd4b48c51f04cc1a6"},"schema_version":"1.0","source":{"id":"2606.07451","kind":"arxiv","version":1}},"canonical_sha256":"bf3670f08aac090956216610ee0fb6fb084837d24d8ee3eba2a288d0c0a4a227","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"bf3670f08aac090956216610ee0fb6fb084837d24d8ee3eba2a288d0c0a4a227","first_computed_at":"2026-06-08T01:05:27.834858Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-08T01:05:27.834858Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"SxnNVVail704xxZ9YfrrWC6SOZBYyKrwSV+SZKNiZNPZhmNLUpXs053+/wSO+TJz6Rm++SkUVUz1vqbB9HWmBg==","signature_status":"signed_v1","signed_at":"2026-06-08T01:05:27.835653Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.07451","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7f0de5be742d321b3a45d2111a8dbd823123f87985bbf30520ebaed09435718b","sha256:b9af860e7dfde09a14b39ad74fe9d78897b861fd826a7dec30f821311ea3ae5c"],"state_sha256":"dc03ddec56ead43e7e711e7250a641134d7227fb4b515dc947ce4b3deddd1bbd"}