{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:S66X43R6VHVRXX73QREKVXHVJR","short_pith_number":"pith:S66X43R6","canonical_record":{"source":{"id":"2505.06356","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-05-09T18:01:50Z","cross_cats_sorted":[],"title_canon_sha256":"e861196fafcfdd9dd713afc34995b279bb0f4559786ef22c80496076bfe71638","abstract_canon_sha256":"01f3281bc7a1985164d7e0e10747a15bd2b660eb02cdf69a20296f5c4123a4b8"},"schema_version":"1.0"},"canonical_sha256":"97bd7e6e3ea9eb1bdffb8448aadcf54c53ef243c54d2772c68f9d00375a86745","source":{"kind":"arxiv","id":"2505.06356","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.06356","created_at":"2026-07-05T11:01:07Z"},{"alias_kind":"arxiv_version","alias_value":"2505.06356v1","created_at":"2026-07-05T11:01:07Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.06356","created_at":"2026-07-05T11:01:07Z"},{"alias_kind":"pith_short_12","alias_value":"S66X43R6VHVR","created_at":"2026-07-05T11:01:07Z"},{"alias_kind":"pith_short_16","alias_value":"S66X43R6VHVRXX73","created_at":"2026-07-05T11:01:07Z"},{"alias_kind":"pith_short_8","alias_value":"S66X43R6","created_at":"2026-07-05T11:01:07Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:S66X43R6VHVRXX73QREKVXHVJR","target":"record","payload":{"canonical_record":{"source":{"id":"2505.06356","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-05-09T18:01:50Z","cross_cats_sorted":[],"title_canon_sha256":"e861196fafcfdd9dd713afc34995b279bb0f4559786ef22c80496076bfe71638","abstract_canon_sha256":"01f3281bc7a1985164d7e0e10747a15bd2b660eb02cdf69a20296f5c4123a4b8"},"schema_version":"1.0"},"canonical_sha256":"97bd7e6e3ea9eb1bdffb8448aadcf54c53ef243c54d2772c68f9d00375a86745","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:01:07.805875Z","signature_b64":"zbcAnSApq9HsEC3lB7x384T0bPF5AdZ+/YdKd5nsF/UbbgJKK/62U8Re7bg0EFCp/55ccJjzOR/9YJb+az0fDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"97bd7e6e3ea9eb1bdffb8448aadcf54c53ef243c54d2772c68f9d00375a86745","last_reissued_at":"2026-07-05T11:01:07.805463Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:01:07.805463Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2505.06356","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":"2026-07-05T11:01:07Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ntiYBLp0Ut/J1+Oa/W/DcihBEAE4zRhe2Ad6k/L3U0J8ML6X1LJn3AW+orPchbsdj7t1NzHYhN5eD6pOW3AkBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-19T04:40:26.536073Z"},"content_sha256":"8c8d88fadaea5c27e758a26bd12957a5320e2a90025ae75b9e49140655b28af5","schema_version":"1.0","event_id":"sha256:8c8d88fadaea5c27e758a26bd12957a5320e2a90025ae75b9e49140655b28af5"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:S66X43R6VHVRXX73QREKVXHVJR","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Understanding and Mitigating Toxicity in Image-Text Pretraining Datasets: A Case Study on LLaVA","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Karthik Reddy Kanjula, Nahid Alam, Shayekh Bin Islam, Surya Guthikonda","submitted_at":"2025-05-09T18:01:50Z","abstract_excerpt":"Pretraining datasets are foundational to the development of multimodal models, yet they often have inherent biases and toxic content from the web-scale corpora they are sourced from. In this paper, we investigate the prevalence of toxicity in LLaVA image-text pretraining dataset, examining how harmful content manifests in different modalities. We present a comprehensive analysis of common toxicity categories and propose targeted mitigation strategies, resulting in the creation of a refined toxicity-mitigated dataset. This dataset removes 7,531 of toxic image-text pairs in the LLaVA pre-trainin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.06356","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/2505.06356/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":"2026-07-05T11:01:07Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"658ePY6MfA6h6csm+3j5/K94YoWz6B970U+PVVVG7W4Blv1IKKkAOtn4LJGNDkI/ftzo4hiP/f4WZwRpJ2YMAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-19T04:40:26.536430Z"},"content_sha256":"055d2007fb48c99f504af47c1df76b19ab2835ce626fcfebb92d2b149a51d767","schema_version":"1.0","event_id":"sha256:055d2007fb48c99f504af47c1df76b19ab2835ce626fcfebb92d2b149a51d767"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/S66X43R6VHVRXX73QREKVXHVJR/bundle.json","state_url":"https://pith.science/pith/S66X43R6VHVRXX73QREKVXHVJR/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/S66X43R6VHVRXX73QREKVXHVJR/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":"2026-08-19T04:40:26Z","links":{"resolver":"https://pith.science/pith/S66X43R6VHVRXX73QREKVXHVJR","bundle":"https://pith.science/pith/S66X43R6VHVRXX73QREKVXHVJR/bundle.json","state":"https://pith.science/pith/S66X43R6VHVRXX73QREKVXHVJR/state.json","well_known_bundle":"https://pith.science/.well-known/pith/S66X43R6VHVRXX73QREKVXHVJR/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:S66X43R6VHVRXX73QREKVXHVJR","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":"01f3281bc7a1985164d7e0e10747a15bd2b660eb02cdf69a20296f5c4123a4b8","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-05-09T18:01:50Z","title_canon_sha256":"e861196fafcfdd9dd713afc34995b279bb0f4559786ef22c80496076bfe71638"},"schema_version":"1.0","source":{"id":"2505.06356","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.06356","created_at":"2026-07-05T11:01:07Z"},{"alias_kind":"arxiv_version","alias_value":"2505.06356v1","created_at":"2026-07-05T11:01:07Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.06356","created_at":"2026-07-05T11:01:07Z"},{"alias_kind":"pith_short_12","alias_value":"S66X43R6VHVR","created_at":"2026-07-05T11:01:07Z"},{"alias_kind":"pith_short_16","alias_value":"S66X43R6VHVRXX73","created_at":"2026-07-05T11:01:07Z"},{"alias_kind":"pith_short_8","alias_value":"S66X43R6","created_at":"2026-07-05T11:01:07Z"}],"graph_snapshots":[{"event_id":"sha256:055d2007fb48c99f504af47c1df76b19ab2835ce626fcfebb92d2b149a51d767","target":"graph","created_at":"2026-07-05T11:01:07Z","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/2505.06356/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Pretraining datasets are foundational to the development of multimodal models, yet they often have inherent biases and toxic content from the web-scale corpora they are sourced from. In this paper, we investigate the prevalence of toxicity in LLaVA image-text pretraining dataset, examining how harmful content manifests in different modalities. We present a comprehensive analysis of common toxicity categories and propose targeted mitigation strategies, resulting in the creation of a refined toxicity-mitigated dataset. This dataset removes 7,531 of toxic image-text pairs in the LLaVA pre-trainin","authors_text":"Karthik Reddy Kanjula, Nahid Alam, Shayekh Bin Islam, Surya Guthikonda","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-05-09T18:01:50Z","title":"Understanding and Mitigating Toxicity in Image-Text Pretraining Datasets: A Case Study on LLaVA"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.06356","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":"sha256:8c8d88fadaea5c27e758a26bd12957a5320e2a90025ae75b9e49140655b28af5","target":"record","created_at":"2026-07-05T11:01:07Z","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":"01f3281bc7a1985164d7e0e10747a15bd2b660eb02cdf69a20296f5c4123a4b8","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-05-09T18:01:50Z","title_canon_sha256":"e861196fafcfdd9dd713afc34995b279bb0f4559786ef22c80496076bfe71638"},"schema_version":"1.0","source":{"id":"2505.06356","kind":"arxiv","version":1}},"canonical_sha256":"97bd7e6e3ea9eb1bdffb8448aadcf54c53ef243c54d2772c68f9d00375a86745","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"97bd7e6e3ea9eb1bdffb8448aadcf54c53ef243c54d2772c68f9d00375a86745","first_computed_at":"2026-07-05T11:01:07.805463Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:01:07.805463Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"zbcAnSApq9HsEC3lB7x384T0bPF5AdZ+/YdKd5nsF/UbbgJKK/62U8Re7bg0EFCp/55ccJjzOR/9YJb+az0fDw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:01:07.805875Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.06356","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:8c8d88fadaea5c27e758a26bd12957a5320e2a90025ae75b9e49140655b28af5","sha256:055d2007fb48c99f504af47c1df76b19ab2835ce626fcfebb92d2b149a51d767"],"state_sha256":"83263ca050c4d71784e2c7493704b67611312a7955e5434d8fcca597ac16989d"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"akVu0/EuCgLUpCNKqF47KTXGEpet4J3+FGgjRyIXRemsCyI4vl4+CH+aIxqHHbXZbIhaX1VAQyqGW123d3q4Cg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-19T04:40:26.539741Z","bundle_sha256":"e5f1747165d07d855a4a2c556a25bc567089603aa16998a0164ad39a712f1d56"}}