{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:DOC2W33DYIXD2HMVKQ2JH6URI4","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":"ffae18e1b962fb77c3230abd25ca467a39cec0e27ba076f53d55e1817feea78d","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IR","submitted_at":"2026-04-23T10:13:21Z","title_canon_sha256":"7809308bf8450041565e2d0a7719fe8b061d5246c5ff381f71c54341c0780830"},"schema_version":"1.0","source":{"id":"2604.21511","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.21511","created_at":"2026-06-02T02:04:17Z"},{"alias_kind":"arxiv_version","alias_value":"2604.21511v2","created_at":"2026-06-02T02:04:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.21511","created_at":"2026-06-02T02:04:17Z"},{"alias_kind":"pith_short_12","alias_value":"DOC2W33DYIXD","created_at":"2026-06-02T02:04:17Z"},{"alias_kind":"pith_short_16","alias_value":"DOC2W33DYIXD2HMV","created_at":"2026-06-02T02:04:17Z"},{"alias_kind":"pith_short_8","alias_value":"DOC2W33D","created_at":"2026-06-02T02:04:17Z"}],"graph_snapshots":[{"event_id":"sha256:2e907cbade080cf5a6f294ebd8955a3d1313877ae5b9b126aff65c053cfc648d","target":"graph","created_at":"2026-06-02T02:04:17Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Our experiments demonstrate that SAE-SPLADE achieves retrieval performance comparable to SPLADE on both in-domain and out-of-domain tasks while offering improved efficiency."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the SAE-learned latent concepts form a drop-in replacement for the original token vocabulary without losing retrieval-critical information and that the two training regimes remain compatible."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"SAE-SPLADE substitutes SPLADE's backbone vocabulary with SAE-derived semantic concepts and matches standard SPLADE performance with better efficiency on in- and out-of-domain tasks."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Replacing the token vocabulary in SPLADE with semantic concepts from Sparse Auto-Encoders yields comparable retrieval performance with improved efficiency."}],"snapshot_sha256":"e4bf3c4da65a7997001e3d16f869f0941c1e86bb6c883182365d10d344cc2bd4"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-21T12:39:54.431790Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-20T00:56:40.743827Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2604.21511/integrity.json","findings":[],"snapshot_sha256":"91559a2c424ad4746806fa56964f845e6a8d759b1e0ac150f6a3077573b0fbbb","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Learned Sparse IR models, such as SPLADE, offer an excellent efficiency-effectiveness tradeoff. However, they rely on the underlying backbone vocabulary, which might hinder performance (polysemicity and synonymy) and pose a challenge for multi-lingual and multi-modal usages. To solve this limitation, we propose to replace the backbone vocabulary with a latent space of semantic concepts learned using Sparse Auto-Encoders (SAE). Throughout this paper, we study the compatibility of these 2 concepts, explore training approaches, and analyze the differences between our SAE-SPLADE model and traditio","authors_text":"Basile Van Cooten, Benjamin Piwowarski, Laure Soulier, Mathias Vast, Yuxuan Zong","cross_cats":["cs.CL"],"headline":"Replacing the token vocabulary in SPLADE with semantic concepts from Sparse Auto-Encoders yields comparable retrieval performance with improved efficiency.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IR","submitted_at":"2026-04-23T10:13:21Z","title":"From Tokens to Concepts: Leveraging SAE for SPLADE"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.21511","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-09T20:26:30.758694Z","id":"88771eaa-bb96-4963-9333-6be14991a1b1","model_set":{"reader":"grok-4.3"},"one_line_summary":"SAE-SPLADE substitutes SPLADE's backbone vocabulary with SAE-derived semantic concepts and matches standard SPLADE performance with better efficiency on in- and out-of-domain tasks.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Replacing the token vocabulary in SPLADE with semantic concepts from Sparse Auto-Encoders yields comparable retrieval performance with improved efficiency.","strongest_claim":"Our experiments demonstrate that SAE-SPLADE achieves retrieval performance comparable to SPLADE on both in-domain and out-of-domain tasks while offering improved efficiency.","weakest_assumption":"That the SAE-learned latent concepts form a drop-in replacement for the original token vocabulary without losing retrieval-critical information and that the two training regimes remain compatible."}},"verdict_id":"88771eaa-bb96-4963-9333-6be14991a1b1"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:0d0a2c895414da0ddcbc714fc98bef840661e71084d879fe671610f2100d6a48","target":"record","created_at":"2026-06-02T02:04:17Z","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":"ffae18e1b962fb77c3230abd25ca467a39cec0e27ba076f53d55e1817feea78d","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IR","submitted_at":"2026-04-23T10:13:21Z","title_canon_sha256":"7809308bf8450041565e2d0a7719fe8b061d5246c5ff381f71c54341c0780830"},"schema_version":"1.0","source":{"id":"2604.21511","kind":"arxiv","version":2}},"canonical_sha256":"1b85ab6f63c22e3d1d95543493fa914733aa2d06286a87aea8c6089003d1160a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1b85ab6f63c22e3d1d95543493fa914733aa2d06286a87aea8c6089003d1160a","first_computed_at":"2026-06-02T02:04:17.954944Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-02T02:04:17.954944Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"w6k0FyzL9xwYD3GRu9E34AnN/+pzId6PlYu/Xh79ewTy8HGU3aviG0atoyLcoTpdcNT6U4Ovx8icVgCT15HLDQ==","signature_status":"signed_v1","signed_at":"2026-06-02T02:04:17.955363Z","signed_message":"canonical_sha256_bytes"},"source_id":"2604.21511","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0d0a2c895414da0ddcbc714fc98bef840661e71084d879fe671610f2100d6a48","sha256:2e907cbade080cf5a6f294ebd8955a3d1313877ae5b9b126aff65c053cfc648d"],"state_sha256":"f5e7a0396dafcc100aa6849ccafd53a6e252f1dc22e70be88014ccfd36b5eb36"}