{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:72ACJEDMQRHOO2JVCUDG2M2WZK","short_pith_number":"pith:72ACJEDM","canonical_record":{"source":{"id":"2605.15961","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-15T13:54:21Z","cross_cats_sorted":[],"title_canon_sha256":"a035f534f25ed218da3145e5da2d20bf8eb021b44527f8ff1104ce259388294e","abstract_canon_sha256":"57b65d836c5cc4f2deebd6b7aa3bcfc98410f4b540c53c611fbc7751b16eff1a"},"schema_version":"1.0"},"canonical_sha256":"fe8024906c844ee7693515066d3356caa677d1f5fae77070f9e460f3c1f4a436","source":{"kind":"arxiv","id":"2605.15961","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.15961","created_at":"2026-05-20T00:01:46Z"},{"alias_kind":"arxiv_version","alias_value":"2605.15961v1","created_at":"2026-05-20T00:01:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15961","created_at":"2026-05-20T00:01:46Z"},{"alias_kind":"pith_short_12","alias_value":"72ACJEDMQRHO","created_at":"2026-05-20T00:01:46Z"},{"alias_kind":"pith_short_16","alias_value":"72ACJEDMQRHOO2JV","created_at":"2026-05-20T00:01:46Z"},{"alias_kind":"pith_short_8","alias_value":"72ACJEDM","created_at":"2026-05-20T00:01:46Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:72ACJEDMQRHOO2JVCUDG2M2WZK","target":"record","payload":{"canonical_record":{"source":{"id":"2605.15961","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-15T13:54:21Z","cross_cats_sorted":[],"title_canon_sha256":"a035f534f25ed218da3145e5da2d20bf8eb021b44527f8ff1104ce259388294e","abstract_canon_sha256":"57b65d836c5cc4f2deebd6b7aa3bcfc98410f4b540c53c611fbc7751b16eff1a"},"schema_version":"1.0"},"canonical_sha256":"fe8024906c844ee7693515066d3356caa677d1f5fae77070f9e460f3c1f4a436","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:01:46.725127Z","signature_b64":"yqjeC2IcJQd/hDmAYSp7tcBZREWklb1BtvCwChKd5Tm9tn2EXCvpoRGPljnrJdq+CTv96Qet90MEgMfC8mKAAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fe8024906c844ee7693515066d3356caa677d1f5fae77070f9e460f3c1f4a436","last_reissued_at":"2026-05-20T00:01:46.724636Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:01:46.724636Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.15961","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":"2026-05-20T00:01:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"0dw//iKUy71RRyDi6D3PuEhWMW5mOhBZe9nrRqy1mj0fruHAtcyXqj9IKKU2rMsqcVHJZXT4ltpV0h1YcuoPBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-25T18:40:13.329058Z"},"content_sha256":"4b8734547c691566257114f7058a12c2e110e27f29f7d230906c977bd0be3e53","schema_version":"1.0","event_id":"sha256:4b8734547c691566257114f7058a12c2e110e27f29f7d230906c977bd0be3e53"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:72ACJEDMQRHOO2JVCUDG2M2WZK","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Sparse Autoencoders enable Robust and Interpretable Fine-tuning of CLIP models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ankit Sonthalia, Arnas Uselis, Fabian Morelli, Seong Joon Oh","submitted_at":"2026-05-15T13:54:21Z","abstract_excerpt":"Large-scale pre-trained vision-language models like CLIP demonstrate remarkable zero-shot performance across diverse tasks. However, fine-tuning these models to improve downstream performance often degrades robustness against distribution shifts. Recent approaches have attempted to mitigate this trade-off, but often rely on computationally expensive text-guidance. We propose a novel method for robust fine-tuning, SAE-FT, which operates only on the model's visual representations. SAE-FT regularizes changes to these representations by penalizing the addition and removal of semantically meaningfu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.15961","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/2605.15961/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-19T17:33:44.875206Z","status":"skipped","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T17:01:55.701666Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"cf4636f586bd86dd0a8edabadd0dfcf46b24b86818a29e8d0d64817ecd2dbefe"},"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":"2026-05-20T00:01:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3XzYZ0Uz+PrttbTCpQANHMR+hBSPBoKrqT87aDKt9jS9Pv4eqLWjV3ee4ts/bnYa4Jbe4F2Z3bGDe6V5SXL+DA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-25T18:40:13.329872Z"},"content_sha256":"3027e875f22f03426b22f83429570c0ddd9a11bb3f1732716af1205590f2bff0","schema_version":"1.0","event_id":"sha256:3027e875f22f03426b22f83429570c0ddd9a11bb3f1732716af1205590f2bff0"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/72ACJEDMQRHOO2JVCUDG2M2WZK/bundle.json","state_url":"https://pith.science/pith/72ACJEDMQRHOO2JVCUDG2M2WZK/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/72ACJEDMQRHOO2JVCUDG2M2WZK/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":"2026-05-25T18:40:13Z","links":{"resolver":"https://pith.science/pith/72ACJEDMQRHOO2JVCUDG2M2WZK","bundle":"https://pith.science/pith/72ACJEDMQRHOO2JVCUDG2M2WZK/bundle.json","state":"https://pith.science/pith/72ACJEDMQRHOO2JVCUDG2M2WZK/state.json","well_known_bundle":"https://pith.science/.well-known/pith/72ACJEDMQRHOO2JVCUDG2M2WZK/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:72ACJEDMQRHOO2JVCUDG2M2WZK","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":"57b65d836c5cc4f2deebd6b7aa3bcfc98410f4b540c53c611fbc7751b16eff1a","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-15T13:54:21Z","title_canon_sha256":"a035f534f25ed218da3145e5da2d20bf8eb021b44527f8ff1104ce259388294e"},"schema_version":"1.0","source":{"id":"2605.15961","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.15961","created_at":"2026-05-20T00:01:46Z"},{"alias_kind":"arxiv_version","alias_value":"2605.15961v1","created_at":"2026-05-20T00:01:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15961","created_at":"2026-05-20T00:01:46Z"},{"alias_kind":"pith_short_12","alias_value":"72ACJEDMQRHO","created_at":"2026-05-20T00:01:46Z"},{"alias_kind":"pith_short_16","alias_value":"72ACJEDMQRHOO2JV","created_at":"2026-05-20T00:01:46Z"},{"alias_kind":"pith_short_8","alias_value":"72ACJEDM","created_at":"2026-05-20T00:01:46Z"}],"graph_snapshots":[{"event_id":"sha256:3027e875f22f03426b22f83429570c0ddd9a11bb3f1732716af1205590f2bff0","target":"graph","created_at":"2026-05-20T00:01:46Z","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":[{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T17:33:44.875206Z","status":"skipped","version":"1.0.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T17:01:55.701666Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2605.15961/integrity.json","findings":[],"snapshot_sha256":"cf4636f586bd86dd0a8edabadd0dfcf46b24b86818a29e8d0d64817ecd2dbefe","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large-scale pre-trained vision-language models like CLIP demonstrate remarkable zero-shot performance across diverse tasks. However, fine-tuning these models to improve downstream performance often degrades robustness against distribution shifts. Recent approaches have attempted to mitigate this trade-off, but often rely on computationally expensive text-guidance. We propose a novel method for robust fine-tuning, SAE-FT, which operates only on the model's visual representations. SAE-FT regularizes changes to these representations by penalizing the addition and removal of semantically meaningfu","authors_text":"Ankit Sonthalia, Arnas Uselis, Fabian Morelli, Seong Joon Oh","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-15T13:54:21Z","title":"Sparse Autoencoders enable Robust and Interpretable Fine-tuning of CLIP models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.15961","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":"sha256:4b8734547c691566257114f7058a12c2e110e27f29f7d230906c977bd0be3e53","target":"record","created_at":"2026-05-20T00:01:46Z","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":"57b65d836c5cc4f2deebd6b7aa3bcfc98410f4b540c53c611fbc7751b16eff1a","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-15T13:54:21Z","title_canon_sha256":"a035f534f25ed218da3145e5da2d20bf8eb021b44527f8ff1104ce259388294e"},"schema_version":"1.0","source":{"id":"2605.15961","kind":"arxiv","version":1}},"canonical_sha256":"fe8024906c844ee7693515066d3356caa677d1f5fae77070f9e460f3c1f4a436","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"fe8024906c844ee7693515066d3356caa677d1f5fae77070f9e460f3c1f4a436","first_computed_at":"2026-05-20T00:01:46.724636Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:01:46.724636Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"yqjeC2IcJQd/hDmAYSp7tcBZREWklb1BtvCwChKd5Tm9tn2EXCvpoRGPljnrJdq+CTv96Qet90MEgMfC8mKAAg==","signature_status":"signed_v1","signed_at":"2026-05-20T00:01:46.725127Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.15961","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4b8734547c691566257114f7058a12c2e110e27f29f7d230906c977bd0be3e53","sha256:3027e875f22f03426b22f83429570c0ddd9a11bb3f1732716af1205590f2bff0"],"state_sha256":"38dc8b01c7890692126a6ff8288ed0ec24fa0f27ef614de415c3026926a6ef2a"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"R1wEmB/dni/3LN7VZIRJX+mWLZ93QNThczJqTvJgFqEZIetiZFJjHcWpWgOQ9ihZKcDb5I+X8QM9CHSugzpICg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-25T18:40:13.332664Z","bundle_sha256":"dfd8713ee423d0ee4480dbe80167b6b3d8ae09deb1070cb20fb2d064ef63d77e"}}