{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:EH35QPE3SBIRXE2RUARZV75WMS","short_pith_number":"pith:EH35QPE3","schema_version":"1.0","canonical_sha256":"21f7d83c9b90511b9351a0239affb66480a740eaea5288ddd6de48e8d625c1fe","source":{"kind":"arxiv","id":"2401.11206","version":1},"attestation_state":"computed","paper":{"title":"InferAligner: Inference-Time Alignment for Harmlessness through Cross-Model Guidance","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Botian Jiang, Chenkun Tan, Dong Zhang, Ke Ren, Linyang Li, Pengyu Wang, Xinghao Wang, Xipeng Qiu","submitted_at":"2024-01-20T10:41:03Z","abstract_excerpt":"With the rapid development of large language models (LLMs), they are not only used as general-purpose AI assistants but are also customized through further fine-tuning to meet the requirements of different applications. A pivotal factor in the success of current LLMs is the alignment process. Current alignment methods, such as supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF), focus on training-time alignment and are often complex and cumbersome to implement. Therefore, we develop \\textbf{InferAligner}, a novel inference-time alignment method that utilizes cros"},"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":"2401.11206","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-20T10:41:03Z","cross_cats_sorted":[],"title_canon_sha256":"451c83d35ebf2650ea3a28c9562a695d00332cfc44425e1947c032e5e1ed31ca","abstract_canon_sha256":"7871ab37b960bc3bb790045f0f6c862d302caa8683d7424c2abcb41d6a7ba9fc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:35:47.276996Z","signature_b64":"LcbvK+wHOwfU8sj7+/VqwkBykROXIgTuLFRsEdNrYSWMPm3a2oxL92Cx6z5+nNFwXxnAO/gUXEA22qUUPQIeCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"21f7d83c9b90511b9351a0239affb66480a740eaea5288ddd6de48e8d625c1fe","last_reissued_at":"2026-07-05T07:35:47.276545Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:35:47.276545Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"InferAligner: Inference-Time Alignment for Harmlessness through Cross-Model Guidance","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Botian Jiang, Chenkun Tan, Dong Zhang, Ke Ren, Linyang Li, Pengyu Wang, Xinghao Wang, Xipeng Qiu","submitted_at":"2024-01-20T10:41:03Z","abstract_excerpt":"With the rapid development of large language models (LLMs), they are not only used as general-purpose AI assistants but are also customized through further fine-tuning to meet the requirements of different applications. A pivotal factor in the success of current LLMs is the alignment process. Current alignment methods, such as supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF), focus on training-time alignment and are often complex and cumbersome to implement. Therefore, we develop \\textbf{InferAligner}, a novel inference-time alignment method that utilizes cros"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.11206","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/2401.11206/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":"2401.11206","created_at":"2026-07-05T07:35:47.276611+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.11206v1","created_at":"2026-07-05T07:35:47.276611+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.11206","created_at":"2026-07-05T07:35:47.276611+00:00"},{"alias_kind":"pith_short_12","alias_value":"EH35QPE3SBIR","created_at":"2026-07-05T07:35:47.276611+00:00"},{"alias_kind":"pith_short_16","alias_value":"EH35QPE3SBIRXE2R","created_at":"2026-07-05T07:35:47.276611+00:00"},{"alias_kind":"pith_short_8","alias_value":"EH35QPE3","created_at":"2026-07-05T07:35:47.276611+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.15030","citing_title":"WARD: Adversarially Robust Defense of Web Agents Against Prompt Injections","ref_index":60,"is_internal_anchor":false},{"citing_arxiv_id":"2604.15789","citing_title":"A Systematic Study of Training-Free Methods for Trustworthy Large Language Models","ref_index":50,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EH35QPE3SBIRXE2RUARZV75WMS","json":"https://pith.science/pith/EH35QPE3SBIRXE2RUARZV75WMS.json","graph_json":"https://pith.science/api/pith-number/EH35QPE3SBIRXE2RUARZV75WMS/graph.json","events_json":"https://pith.science/api/pith-number/EH35QPE3SBIRXE2RUARZV75WMS/events.json","paper":"https://pith.science/paper/EH35QPE3"},"agent_actions":{"view_html":"https://pith.science/pith/EH35QPE3SBIRXE2RUARZV75WMS","download_json":"https://pith.science/pith/EH35QPE3SBIRXE2RUARZV75WMS.json","view_paper":"https://pith.science/paper/EH35QPE3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.11206&json=true","fetch_graph":"https://pith.science/api/pith-number/EH35QPE3SBIRXE2RUARZV75WMS/graph.json","fetch_events":"https://pith.science/api/pith-number/EH35QPE3SBIRXE2RUARZV75WMS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EH35QPE3SBIRXE2RUARZV75WMS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EH35QPE3SBIRXE2RUARZV75WMS/action/storage_attestation","attest_author":"https://pith.science/pith/EH35QPE3SBIRXE2RUARZV75WMS/action/author_attestation","sign_citation":"https://pith.science/pith/EH35QPE3SBIRXE2RUARZV75WMS/action/citation_signature","submit_replication":"https://pith.science/pith/EH35QPE3SBIRXE2RUARZV75WMS/action/replication_record"}},"created_at":"2026-07-05T07:35:47.276611+00:00","updated_at":"2026-07-05T07:35:47.276611+00:00"}