{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:BCPIXLYJUECWJGFEWC5RRA3G4N","short_pith_number":"pith:BCPIXLYJ","schema_version":"1.0","canonical_sha256":"089e8baf09a1056498a4b0bb188366e35d068d76d251a63cfc9109b83e193317","source":{"kind":"arxiv","id":"2601.03089","version":2},"attestation_state":"computed","paper":{"title":"Faithfulness Evaluation for Decoder-only LLM Attributions with Controlled Retained Information","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Antoni B. Chan, Xin Huang","submitted_at":"2026-01-06T15:22:39Z","abstract_excerpt":"Large Language Models (LLMs) are increasingly evaluated with input attribution methods, yet comparing such explanations remains challenging. Existing soft-perturbation faithfulness metrics, such as Soft-NC and Soft-NS, can conflate attribution quality with the number of words retained during perturbation: attribution methods with larger average scores may keep more words and therefore obtain inflated scores. To address this issue, we propose $\\pi$-Soft-NC and $\\pi$-Soft-NS, an evaluation framework that compares attribution methods under the same expected retaining probability, thus controlling"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2601.03089","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-01-06T15:22:39Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"799239681cd8280de35c163416a28811858a5360676311785816864df593481e","abstract_canon_sha256":"0fa8d1d9a6000ccd9a0f1a814fcf8eac55f92eeffb5f047b989e39c7352dd783"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-27T02:05:10.150683Z","signature_b64":"1s/OI2HoXjBk34fTnlGkNREAxaOsaQQhRI4WKygJ4wBAP9Yud1RLl98xZawOAD3zgtMZF0opM8/3kAAkcWRaBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"089e8baf09a1056498a4b0bb188366e35d068d76d251a63cfc9109b83e193317","last_reissued_at":"2026-05-27T02:05:10.149801Z","signature_status":"signed_v1","first_computed_at":"2026-05-27T02:05:10.149801Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Faithfulness Evaluation for Decoder-only LLM Attributions with Controlled Retained Information","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Antoni B. Chan, Xin Huang","submitted_at":"2026-01-06T15:22:39Z","abstract_excerpt":"Large Language Models (LLMs) are increasingly evaluated with input attribution methods, yet comparing such explanations remains challenging. Existing soft-perturbation faithfulness metrics, such as Soft-NC and Soft-NS, can conflate attribution quality with the number of words retained during perturbation: attribution methods with larger average scores may keep more words and therefore obtain inflated scores. To address this issue, we propose $\\pi$-Soft-NC and $\\pi$-Soft-NS, an evaluation framework that compares attribution methods under the same expected retaining probability, thus controlling"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.03089","kind":"arxiv","version":2},"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/2601.03089/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2601.03089","created_at":"2026-05-27T02:05:10.149908+00:00"},{"alias_kind":"arxiv_version","alias_value":"2601.03089v2","created_at":"2026-05-27T02:05:10.149908+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.03089","created_at":"2026-05-27T02:05:10.149908+00:00"},{"alias_kind":"pith_short_12","alias_value":"BCPIXLYJUECW","created_at":"2026-05-27T02:05:10.149908+00:00"},{"alias_kind":"pith_short_16","alias_value":"BCPIXLYJUECWJGFE","created_at":"2026-05-27T02:05:10.149908+00:00"},{"alias_kind":"pith_short_8","alias_value":"BCPIXLYJ","created_at":"2026-05-27T02:05:10.149908+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BCPIXLYJUECWJGFEWC5RRA3G4N","json":"https://pith.science/pith/BCPIXLYJUECWJGFEWC5RRA3G4N.json","graph_json":"https://pith.science/api/pith-number/BCPIXLYJUECWJGFEWC5RRA3G4N/graph.json","events_json":"https://pith.science/api/pith-number/BCPIXLYJUECWJGFEWC5RRA3G4N/events.json","paper":"https://pith.science/paper/BCPIXLYJ"},"agent_actions":{"view_html":"https://pith.science/pith/BCPIXLYJUECWJGFEWC5RRA3G4N","download_json":"https://pith.science/pith/BCPIXLYJUECWJGFEWC5RRA3G4N.json","view_paper":"https://pith.science/paper/BCPIXLYJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2601.03089&json=true","fetch_graph":"https://pith.science/api/pith-number/BCPIXLYJUECWJGFEWC5RRA3G4N/graph.json","fetch_events":"https://pith.science/api/pith-number/BCPIXLYJUECWJGFEWC5RRA3G4N/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BCPIXLYJUECWJGFEWC5RRA3G4N/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BCPIXLYJUECWJGFEWC5RRA3G4N/action/storage_attestation","attest_author":"https://pith.science/pith/BCPIXLYJUECWJGFEWC5RRA3G4N/action/author_attestation","sign_citation":"https://pith.science/pith/BCPIXLYJUECWJGFEWC5RRA3G4N/action/citation_signature","submit_replication":"https://pith.science/pith/BCPIXLYJUECWJGFEWC5RRA3G4N/action/replication_record"}},"created_at":"2026-05-27T02:05:10.149908+00:00","updated_at":"2026-05-27T02:05:10.149908+00:00"}