{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:DXP63YCJH66Z5KFW25IVAVDCDW","short_pith_number":"pith:DXP63YCJ","canonical_record":{"source":{"id":"2606.26036","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-24T17:12:42Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"bf971a0e67d3d122506ba885c3af33cd78311a6590b906be03c71985f0ff9156","abstract_canon_sha256":"160be7ea604ab1a0c5afea3ed8a295c78877e2827627c2eb0418429426a23a3f"},"schema_version":"1.0"},"canonical_sha256":"1ddfede0493fbd9ea8b6d7515054621da790322ba15356afe54c94cac6f876f6","source":{"kind":"arxiv","id":"2606.26036","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.26036","created_at":"2026-06-25T01:18:46Z"},{"alias_kind":"arxiv_version","alias_value":"2606.26036v1","created_at":"2026-06-25T01:18:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.26036","created_at":"2026-06-25T01:18:46Z"},{"alias_kind":"pith_short_12","alias_value":"DXP63YCJH66Z","created_at":"2026-06-25T01:18:46Z"},{"alias_kind":"pith_short_16","alias_value":"DXP63YCJH66Z5KFW","created_at":"2026-06-25T01:18:46Z"},{"alias_kind":"pith_short_8","alias_value":"DXP63YCJ","created_at":"2026-06-25T01:18:46Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:DXP63YCJH66Z5KFW25IVAVDCDW","target":"record","payload":{"canonical_record":{"source":{"id":"2606.26036","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-24T17:12:42Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"bf971a0e67d3d122506ba885c3af33cd78311a6590b906be03c71985f0ff9156","abstract_canon_sha256":"160be7ea604ab1a0c5afea3ed8a295c78877e2827627c2eb0418429426a23a3f"},"schema_version":"1.0"},"canonical_sha256":"1ddfede0493fbd9ea8b6d7515054621da790322ba15356afe54c94cac6f876f6","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-25T01:18:46.453349Z","signature_b64":"C6nemSvalrob01fl28RzSnWERAzRtLkrSx1l8KUxgJy1zrWn2YmyfgMIS4ROB5brArhflsU7gwG12scxjC+tAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1ddfede0493fbd9ea8b6d7515054621da790322ba15356afe54c94cac6f876f6","last_reissued_at":"2026-06-25T01:18:46.452935Z","signature_status":"signed_v1","first_computed_at":"2026-06-25T01:18:46.452935Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2606.26036","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":"2026-06-25T01:18:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QGQp4OmU+AMtFOBiEen+FpGyO/75krCj/1PMMcPbxutcHCmm0SWTHWnEvczfj12lobxfUlcIAvnAWP2j3GZTCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-30T04:58:40.529603Z"},"content_sha256":"1fa9ade3875396a0361f1160b00fedcc57da282fc7597d3c3885d74c13606ba2","schema_version":"1.0","event_id":"sha256:1fa9ade3875396a0361f1160b00fedcc57da282fc7597d3c3885d74c13606ba2"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:DXP63YCJH66Z5KFW25IVAVDCDW","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Detect, Unlearn, Restore: Defending Text Summarization Models Against Data Poisoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.CL","authors_text":"Poojitha Thota, Shirin Nilizadeh","submitted_at":"2026-06-24T17:12:42Z","abstract_excerpt":"Training-time data poisoning during fine-tuning poses a significant threat to large language models (LLMs) deployed for abstractive text summarization, where small task-specific datasets exert disproportionate influence on model behavior. In this setting, adversaries manipulate fine-tuning data to induce persistent summarization failures, such as biased or harmful summaries, while preserving standard evaluation metrics. We present a unified post-hoc defense framework for detecting and remediating fine-tuning-stage poisoning in summarization models across the machine learning supply chain. Our "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.26036","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/2606.26036/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":"2026-06-25T01:18:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"VqLMBydKE6D5I/Q7Ke8kDp6rlfH9mAwNOwSVy/orT8C7vR6vDUUFsl1g85nFaUojlsjX44UxCd8T+zxaB0mfBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-30T04:58:40.529971Z"},"content_sha256":"8b3024427617dcf62888d13dd93d5c236528a63fc30edd15e14de99b9aa08291","schema_version":"1.0","event_id":"sha256:8b3024427617dcf62888d13dd93d5c236528a63fc30edd15e14de99b9aa08291"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DXP63YCJH66Z5KFW25IVAVDCDW/bundle.json","state_url":"https://pith.science/pith/DXP63YCJH66Z5KFW25IVAVDCDW/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DXP63YCJH66Z5KFW25IVAVDCDW/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":"2026-06-30T04:58:40Z","links":{"resolver":"https://pith.science/pith/DXP63YCJH66Z5KFW25IVAVDCDW","bundle":"https://pith.science/pith/DXP63YCJH66Z5KFW25IVAVDCDW/bundle.json","state":"https://pith.science/pith/DXP63YCJH66Z5KFW25IVAVDCDW/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DXP63YCJH66Z5KFW25IVAVDCDW/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:DXP63YCJH66Z5KFW25IVAVDCDW","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":"160be7ea604ab1a0c5afea3ed8a295c78877e2827627c2eb0418429426a23a3f","cross_cats_sorted":["cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-24T17:12:42Z","title_canon_sha256":"bf971a0e67d3d122506ba885c3af33cd78311a6590b906be03c71985f0ff9156"},"schema_version":"1.0","source":{"id":"2606.26036","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.26036","created_at":"2026-06-25T01:18:46Z"},{"alias_kind":"arxiv_version","alias_value":"2606.26036v1","created_at":"2026-06-25T01:18:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.26036","created_at":"2026-06-25T01:18:46Z"},{"alias_kind":"pith_short_12","alias_value":"DXP63YCJH66Z","created_at":"2026-06-25T01:18:46Z"},{"alias_kind":"pith_short_16","alias_value":"DXP63YCJH66Z5KFW","created_at":"2026-06-25T01:18:46Z"},{"alias_kind":"pith_short_8","alias_value":"DXP63YCJ","created_at":"2026-06-25T01:18:46Z"}],"graph_snapshots":[{"event_id":"sha256:8b3024427617dcf62888d13dd93d5c236528a63fc30edd15e14de99b9aa08291","target":"graph","created_at":"2026-06-25T01:18:46Z","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":[],"endpoint":"/pith/2606.26036/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Training-time data poisoning during fine-tuning poses a significant threat to large language models (LLMs) deployed for abstractive text summarization, where small task-specific datasets exert disproportionate influence on model behavior. In this setting, adversaries manipulate fine-tuning data to induce persistent summarization failures, such as biased or harmful summaries, while preserving standard evaluation metrics. We present a unified post-hoc defense framework for detecting and remediating fine-tuning-stage poisoning in summarization models across the machine learning supply chain. Our ","authors_text":"Poojitha Thota, Shirin Nilizadeh","cross_cats":["cs.CR"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-24T17:12:42Z","title":"Detect, Unlearn, Restore: Defending Text Summarization Models Against Data Poisoning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.26036","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":"sha256:1fa9ade3875396a0361f1160b00fedcc57da282fc7597d3c3885d74c13606ba2","target":"record","created_at":"2026-06-25T01:18:46Z","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":"160be7ea604ab1a0c5afea3ed8a295c78877e2827627c2eb0418429426a23a3f","cross_cats_sorted":["cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-24T17:12:42Z","title_canon_sha256":"bf971a0e67d3d122506ba885c3af33cd78311a6590b906be03c71985f0ff9156"},"schema_version":"1.0","source":{"id":"2606.26036","kind":"arxiv","version":1}},"canonical_sha256":"1ddfede0493fbd9ea8b6d7515054621da790322ba15356afe54c94cac6f876f6","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1ddfede0493fbd9ea8b6d7515054621da790322ba15356afe54c94cac6f876f6","first_computed_at":"2026-06-25T01:18:46.452935Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-25T01:18:46.452935Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"C6nemSvalrob01fl28RzSnWERAzRtLkrSx1l8KUxgJy1zrWn2YmyfgMIS4ROB5brArhflsU7gwG12scxjC+tAA==","signature_status":"signed_v1","signed_at":"2026-06-25T01:18:46.453349Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.26036","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1fa9ade3875396a0361f1160b00fedcc57da282fc7597d3c3885d74c13606ba2","sha256:8b3024427617dcf62888d13dd93d5c236528a63fc30edd15e14de99b9aa08291"],"state_sha256":"65d11e8a4b9074bfa3bdcf02ed3f6b160d3caa57bb144a4dd3f3ce0da6447e8f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/l5nrlUo8OTkE+jJI5mYkx38lg86dsfzO4q7If1bDGoCjGm5Ke2JcYMRyffbTJ9mQ1rc8pRtn38vyVHAe+t2BA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-30T04:58:40.531987Z","bundle_sha256":"991c58aad9711b5881ae05ff72c67a73d50a18fca53bd76aadb2023121dcd4da"}}