{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:7QXHKRGBZ65G2H66P3TE3WSPWT","short_pith_number":"pith:7QXHKRGB","canonical_record":{"source":{"id":"2606.26437","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-24T23:00:09Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"385bd22942dd14128494fa01569a155c625d4630ea19a6936c8eb8667271a294","abstract_canon_sha256":"6663d9cd4757c538e3b07e83841d077e1b62de348185a8fb64496523fa1d9ce2"},"schema_version":"1.0"},"canonical_sha256":"fc2e7544c1cfba6d1fde7ee64dda4fb4ceebd8d8452d1e4dd9996199053dbf39","source":{"kind":"arxiv","id":"2606.26437","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.26437","created_at":"2026-06-26T00:15:41Z"},{"alias_kind":"arxiv_version","alias_value":"2606.26437v1","created_at":"2026-06-26T00:15:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.26437","created_at":"2026-06-26T00:15:41Z"},{"alias_kind":"pith_short_12","alias_value":"7QXHKRGBZ65G","created_at":"2026-06-26T00:15:41Z"},{"alias_kind":"pith_short_16","alias_value":"7QXHKRGBZ65G2H66","created_at":"2026-06-26T00:15:41Z"},{"alias_kind":"pith_short_8","alias_value":"7QXHKRGB","created_at":"2026-06-26T00:15:41Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:7QXHKRGBZ65G2H66P3TE3WSPWT","target":"record","payload":{"canonical_record":{"source":{"id":"2606.26437","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-24T23:00:09Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"385bd22942dd14128494fa01569a155c625d4630ea19a6936c8eb8667271a294","abstract_canon_sha256":"6663d9cd4757c538e3b07e83841d077e1b62de348185a8fb64496523fa1d9ce2"},"schema_version":"1.0"},"canonical_sha256":"fc2e7544c1cfba6d1fde7ee64dda4fb4ceebd8d8452d1e4dd9996199053dbf39","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-26T00:15:41.235287Z","signature_b64":"vKIfih4vcumTLEsuAdZSXo0xSlbXTMvMBAOJIhrm9mBdwqkgMXkf6UjW4huKBrybcm374qbCNCNyrLtIThBBDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fc2e7544c1cfba6d1fde7ee64dda4fb4ceebd8d8452d1e4dd9996199053dbf39","last_reissued_at":"2026-06-26T00:15:41.234878Z","signature_status":"signed_v1","first_computed_at":"2026-06-26T00:15:41.234878Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2606.26437","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":"2026-06-26T00:15:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xwbGuwKlbWd6kZFuFptiALEOoRUpb9uQLtdDgP1q9ZRIMXhUsYm1xIHw7COuskJkOzQk0WDHfJEsbJvY71LZAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-30T00:41:24.935748Z"},"content_sha256":"50d8d77e6e45dfdbe983a918bd3c29d6d57f694a75ed28110415373cbf7d02ba","schema_version":"1.0","event_id":"sha256:50d8d77e6e45dfdbe983a918bd3c29d6d57f694a75ed28110415373cbf7d02ba"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:7QXHKRGBZ65G2H66P3TE3WSPWT","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"ConflictScore: Identifying and Measuring How Language Models Handle Conflicting Evidence","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Aaron Halfaker, Dan Roth, Patrick Xia, Siyi Liu","submitted_at":"2026-06-24T23:00:09Z","abstract_excerpt":"Existing metrics for factuality and faithfulness evaluate whether an answer is supported or contradicted by its grounding documents, but they fail to capture when both supporting and contradicting evidence coexist. We introduce ConflictScore, a novel metric that quantifies how well a model's response acknowledges conflicting evidence in its grounding documents. Our framework decomposes responses into atomic claims, labels each claim against each grounding document, and then aggregates these labels into two complementary measures: ConflictScore-Count (CS-C), the proportion of claims exhibiting "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.26437","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/2606.26437/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":"2026-06-26T00:15:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"PPk2/5LSlc9mvqbD/0F3HZGIvEnzI7s/sdayOEfDemRhrFzyQRAGYWn+bSfb//wpmHYnvJ31QcO0Ae+T6H+9Bg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-30T00:41:24.936135Z"},"content_sha256":"d83e5aebbd42c3cdbdc3dde751823cc616cb3653994a7268e6d578994ea806f2","schema_version":"1.0","event_id":"sha256:d83e5aebbd42c3cdbdc3dde751823cc616cb3653994a7268e6d578994ea806f2"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/7QXHKRGBZ65G2H66P3TE3WSPWT/bundle.json","state_url":"https://pith.science/pith/7QXHKRGBZ65G2H66P3TE3WSPWT/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/7QXHKRGBZ65G2H66P3TE3WSPWT/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":"2026-06-30T00:41:24Z","links":{"resolver":"https://pith.science/pith/7QXHKRGBZ65G2H66P3TE3WSPWT","bundle":"https://pith.science/pith/7QXHKRGBZ65G2H66P3TE3WSPWT/bundle.json","state":"https://pith.science/pith/7QXHKRGBZ65G2H66P3TE3WSPWT/state.json","well_known_bundle":"https://pith.science/.well-known/pith/7QXHKRGBZ65G2H66P3TE3WSPWT/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:7QXHKRGBZ65G2H66P3TE3WSPWT","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":"6663d9cd4757c538e3b07e83841d077e1b62de348185a8fb64496523fa1d9ce2","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-24T23:00:09Z","title_canon_sha256":"385bd22942dd14128494fa01569a155c625d4630ea19a6936c8eb8667271a294"},"schema_version":"1.0","source":{"id":"2606.26437","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.26437","created_at":"2026-06-26T00:15:41Z"},{"alias_kind":"arxiv_version","alias_value":"2606.26437v1","created_at":"2026-06-26T00:15:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.26437","created_at":"2026-06-26T00:15:41Z"},{"alias_kind":"pith_short_12","alias_value":"7QXHKRGBZ65G","created_at":"2026-06-26T00:15:41Z"},{"alias_kind":"pith_short_16","alias_value":"7QXHKRGBZ65G2H66","created_at":"2026-06-26T00:15:41Z"},{"alias_kind":"pith_short_8","alias_value":"7QXHKRGB","created_at":"2026-06-26T00:15:41Z"}],"graph_snapshots":[{"event_id":"sha256:d83e5aebbd42c3cdbdc3dde751823cc616cb3653994a7268e6d578994ea806f2","target":"graph","created_at":"2026-06-26T00:15:41Z","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/2606.26437/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Existing metrics for factuality and faithfulness evaluate whether an answer is supported or contradicted by its grounding documents, but they fail to capture when both supporting and contradicting evidence coexist. We introduce ConflictScore, a novel metric that quantifies how well a model's response acknowledges conflicting evidence in its grounding documents. Our framework decomposes responses into atomic claims, labels each claim against each grounding document, and then aggregates these labels into two complementary measures: ConflictScore-Count (CS-C), the proportion of claims exhibiting ","authors_text":"Aaron Halfaker, Dan Roth, Patrick Xia, Siyi Liu","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-24T23:00:09Z","title":"ConflictScore: Identifying and Measuring How Language Models Handle Conflicting Evidence"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.26437","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":"sha256:50d8d77e6e45dfdbe983a918bd3c29d6d57f694a75ed28110415373cbf7d02ba","target":"record","created_at":"2026-06-26T00:15:41Z","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":"6663d9cd4757c538e3b07e83841d077e1b62de348185a8fb64496523fa1d9ce2","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-24T23:00:09Z","title_canon_sha256":"385bd22942dd14128494fa01569a155c625d4630ea19a6936c8eb8667271a294"},"schema_version":"1.0","source":{"id":"2606.26437","kind":"arxiv","version":1}},"canonical_sha256":"fc2e7544c1cfba6d1fde7ee64dda4fb4ceebd8d8452d1e4dd9996199053dbf39","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"fc2e7544c1cfba6d1fde7ee64dda4fb4ceebd8d8452d1e4dd9996199053dbf39","first_computed_at":"2026-06-26T00:15:41.234878Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-26T00:15:41.234878Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"vKIfih4vcumTLEsuAdZSXo0xSlbXTMvMBAOJIhrm9mBdwqkgMXkf6UjW4huKBrybcm374qbCNCNyrLtIThBBDA==","signature_status":"signed_v1","signed_at":"2026-06-26T00:15:41.235287Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.26437","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:50d8d77e6e45dfdbe983a918bd3c29d6d57f694a75ed28110415373cbf7d02ba","sha256:d83e5aebbd42c3cdbdc3dde751823cc616cb3653994a7268e6d578994ea806f2"],"state_sha256":"522bae15236ef03d24fb1ad1a1b6751431ecbe6ca928a6895b451503674d8935"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"U7sVJoVlSKyW1JdoNwQ4f9A5qx9QrrK9TbAPCS79P6nQRaPS7uLh3mu7KQqq8mW8DsKyDTVU6YR3bDQIlP5GCA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-30T00:41:24.938081Z","bundle_sha256":"a6fd74b52c9089a072d8f72bf01dc3fc46d2f5b0c6a94792409cf27b583bb33c"}}