{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:HSPOK5VN2LCBPFIJXBTSX45JCR","short_pith_number":"pith:HSPOK5VN","schema_version":"1.0","canonical_sha256":"3c9ee576add2c4179509b8672bf3a9145e2eddb17055a71cf5a98061cacbf45a","source":{"kind":"arxiv","id":"2604.23862","version":2},"attestation_state":"computed","paper":{"title":"Graph Memory Transformer (GMT)","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A learned graph of memory centroids can replace the feed-forward sublayer in a decoder-only transformer.","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Niccol\\`o Ferrari, Nicola Zanarini","submitted_at":"2026-04-26T20:09:25Z","abstract_excerpt":"We investigate whether the Feed-Forward Network (FFN) sublayer in a decoder-only transformer can be replaced by an explicit learned memory graph while preserving the surrounding autoregressive architecture. The proposed Graph Memory Transformer (GMT) keeps causal self-attention intact, but replaces the usual per-token FFN transformation with a memory cell that routes token representations over a learned bank of centroids connected by a learned directed transition matrix. In the base GMT v7 instantiation studied here, each of 16 transformer blocks contains 128 centroids, a 128 * 128 edge matrix"},"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":"2604.23862","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-04-26T20:09:25Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"c974ccea7fca6e2627f929eb263e0832b8eda1dd2c6feaf1cae9d80dc56ee91e","abstract_canon_sha256":"22ae36e7fb6ac610402f580714ab826f9c81da91cb3e88afd7dafb4f274cd42c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T02:04:48.216901Z","signature_b64":"wlrnZGzznPYj7MHRYHQ90bYWE2aBFnpSqyiL6OfbgH+ME6U3ArKVcA9YosLwjINMDrUXv51fgbUKX0FYaroKBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3c9ee576add2c4179509b8672bf3a9145e2eddb17055a71cf5a98061cacbf45a","last_reissued_at":"2026-05-28T02:04:48.216382Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T02:04:48.216382Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Graph Memory Transformer (GMT)","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A learned graph of memory centroids can replace the feed-forward sublayer in a decoder-only transformer.","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Niccol\\`o Ferrari, Nicola Zanarini","submitted_at":"2026-04-26T20:09:25Z","abstract_excerpt":"We investigate whether the Feed-Forward Network (FFN) sublayer in a decoder-only transformer can be replaced by an explicit learned memory graph while preserving the surrounding autoregressive architecture. The proposed Graph Memory Transformer (GMT) keeps causal self-attention intact, but replaces the usual per-token FFN transformation with a memory cell that routes token representations over a learned bank of centroids connected by a learned directed transition matrix. In the base GMT v7 instantiation studied here, each of 16 transformer blocks contains 128 centroids, a 128 * 128 edge matrix"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"The base GMT v7 model trains stably and exposes centroid usage, transition structure, and source-to-target movement as directly inspectable quantities of the forward computation. It remains behind the larger dense baseline in validation loss and perplexity (3.5995/36.58 vs. 3.2903/26.85), while showing close zero-shot benchmark behavior under the evaluated setting.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That gravitational source routing plus token-conditioned target selection plus gated displacement readout can functionally substitute for the dense FFN transformation without requiring additional capacity or architectural changes outside the memory cell itself.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Graph Memory Transformer (GMT) swaps dense FFN sublayers for a graph of 128 centroids and a learned 128x128 transition matrix per block, yielding a 82M-parameter decoder-only LM that trains stably but trails a 103M dense baseline in perplexity.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A learned graph of memory centroids can replace the feed-forward sublayer in a decoder-only transformer.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"622e7f307e16a4a43f6e050b20f4b25425ce585f5d9c7e61d1f9238ecdf6a960"},"source":{"id":"2604.23862","kind":"arxiv","version":2},"verdict":{"id":"66cf2d24-7e13-42cf-bc27-690c491710d1","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-08T06:24:14.512096Z","strongest_claim":"The base GMT v7 model trains stably and exposes centroid usage, transition structure, and source-to-target movement as directly inspectable quantities of the forward computation. It remains behind the larger dense baseline in validation loss and perplexity (3.5995/36.58 vs. 3.2903/26.85), while showing close zero-shot benchmark behavior under the evaluated setting.","one_line_summary":"Graph Memory Transformer (GMT) swaps dense FFN sublayers for a graph of 128 centroids and a learned 128x128 transition matrix per block, yielding a 82M-parameter decoder-only LM that trains stably but trails a 103M dense baseline in perplexity.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That gravitational source routing plus token-conditioned target selection plus gated displacement readout can functionally substitute for the dense FFN transformation without requiring additional capacity or architectural changes outside the memory cell itself.","pith_extraction_headline":"A learned graph of memory centroids can replace the feed-forward sublayer in a decoder-only transformer."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.23862/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-21T07:43:11.410025Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T22:42:36.913523Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"f0922dd35a64782136d1f08b38c0aeb566f919a7180a5d3b6d3866065d4fb412"},"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":"2604.23862","created_at":"2026-05-28T02:04:48.216447+00:00"},{"alias_kind":"arxiv_version","alias_value":"2604.23862v2","created_at":"2026-05-28T02:04:48.216447+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.23862","created_at":"2026-05-28T02:04:48.216447+00:00"},{"alias_kind":"pith_short_12","alias_value":"HSPOK5VN2LCB","created_at":"2026-05-28T02:04:48.216447+00:00"},{"alias_kind":"pith_short_16","alias_value":"HSPOK5VN2LCBPFIJ","created_at":"2026-05-28T02:04:48.216447+00:00"},{"alias_kind":"pith_short_8","alias_value":"HSPOK5VN","created_at":"2026-05-28T02:04:48.216447+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/HSPOK5VN2LCBPFIJXBTSX45JCR","json":"https://pith.science/pith/HSPOK5VN2LCBPFIJXBTSX45JCR.json","graph_json":"https://pith.science/api/pith-number/HSPOK5VN2LCBPFIJXBTSX45JCR/graph.json","events_json":"https://pith.science/api/pith-number/HSPOK5VN2LCBPFIJXBTSX45JCR/events.json","paper":"https://pith.science/paper/HSPOK5VN"},"agent_actions":{"view_html":"https://pith.science/pith/HSPOK5VN2LCBPFIJXBTSX45JCR","download_json":"https://pith.science/pith/HSPOK5VN2LCBPFIJXBTSX45JCR.json","view_paper":"https://pith.science/paper/HSPOK5VN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2604.23862&json=true","fetch_graph":"https://pith.science/api/pith-number/HSPOK5VN2LCBPFIJXBTSX45JCR/graph.json","fetch_events":"https://pith.science/api/pith-number/HSPOK5VN2LCBPFIJXBTSX45JCR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HSPOK5VN2LCBPFIJXBTSX45JCR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HSPOK5VN2LCBPFIJXBTSX45JCR/action/storage_attestation","attest_author":"https://pith.science/pith/HSPOK5VN2LCBPFIJXBTSX45JCR/action/author_attestation","sign_citation":"https://pith.science/pith/HSPOK5VN2LCBPFIJXBTSX45JCR/action/citation_signature","submit_replication":"https://pith.science/pith/HSPOK5VN2LCBPFIJXBTSX45JCR/action/replication_record"}},"created_at":"2026-05-28T02:04:48.216447+00:00","updated_at":"2026-05-28T02:04:48.216447+00:00"}