{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:BMM5EYN7NAZNSW7VKT2LCMP7T4","short_pith_number":"pith:BMM5EYN7","canonical_record":{"source":{"id":"2604.17121","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-18T19:46:30Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"dc9ab62c6a6a71e3f67e75c33518c75c8f4a3620545822a512863e335212cefc","abstract_canon_sha256":"86cc7ddcbfb88ccb7429888589ecae5153cb07ffe5f9b45849786d86c5c71403"},"schema_version":"1.0"},"canonical_sha256":"0b19d261bf6832d95bf554f4b131ff9f3902f55dbceb5e559f14ceb4abf67051","source":{"kind":"arxiv","id":"2604.17121","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.17121","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"arxiv_version","alias_value":"2604.17121v3","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.17121","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_12","alias_value":"BMM5EYN7NAZN","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_16","alias_value":"BMM5EYN7NAZNSW7V","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_8","alias_value":"BMM5EYN7","created_at":"2026-06-05T01:14:38Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:BMM5EYN7NAZNSW7VKT2LCMP7T4","target":"record","payload":{"canonical_record":{"source":{"id":"2604.17121","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-18T19:46:30Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"dc9ab62c6a6a71e3f67e75c33518c75c8f4a3620545822a512863e335212cefc","abstract_canon_sha256":"86cc7ddcbfb88ccb7429888589ecae5153cb07ffe5f9b45849786d86c5c71403"},"schema_version":"1.0"},"canonical_sha256":"0b19d261bf6832d95bf554f4b131ff9f3902f55dbceb5e559f14ceb4abf67051","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-05T01:14:38.908766Z","signature_b64":"YHVGOotFijwRUVe6GZSmi6BC71/T5k+XHjvKYzMX/8P4pk337933dfXT61TlGttuppa8i53XmagzTHaIJmJ2Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0b19d261bf6832d95bf554f4b131ff9f3902f55dbceb5e559f14ceb4abf67051","last_reissued_at":"2026-06-05T01:14:38.907997Z","signature_status":"signed_v1","first_computed_at":"2026-06-05T01:14:38.907997Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2604.17121","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-05T01:14:38Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"0V9vN9gQXi7VriHopQ6vUmPDhUnbX+Ox6ZEaDk0Bm2UqhQo3OLyHHSluJw3A0m/Na673hlrxc+uwzTxSY8WfDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-12T06:53:22.867505Z"},"content_sha256":"413093d936351f110618857d09abc6a502783ffa173b02afec8b93cb319ff45b","schema_version":"1.0","event_id":"sha256:413093d936351f110618857d09abc6a502783ffa173b02afec8b93cb319ff45b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:BMM5EYN7NAZNSW7VKT2LCMP7T4","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"The Topological Trouble With Transformers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Transformers push evolving state representations deeper into their layers with each new input, exhausting depth and limiting dynamic tracking.","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Michael C. Mozer, Rosanne Liu, Shoaib Ahmed Siddiqui","submitted_at":"2026-04-18T19:46:30Z","abstract_excerpt":"Transformers encode structure in sequences via an expanding contextual history. However, their purely feedforward architecture fundamentally limits dynamic state tracking. State tracking -- the iterative updating of latent variables reflecting an evolving environment -- involves inherently sequential dependencies that feedforward networks struggle to maintain. Consequently, feedforward models push evolving state representations deeper into their layer stack with each new input step, rendering information inaccessible in shallow layers and ultimately exhausting the model's depth. While this dep"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"their purely feedforward architecture fundamentally limits dynamic state tracking. State tracking -- the iterative updating of latent variables reflecting an evolving environment -- involves inherently sequential dependencies that feedforward networks struggle to maintain. Consequently, feedforward models push evolving state representations deeper into their layer stack with each new input step, rendering information inaccessible in shallow layers and ultimately exhausting the model's depth.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That dynamic depth models, explicit thinking traces, and latent thinking are inherently too computationally and memory inefficient to serve as scalable solutions, and that recurrent architectures will integrate state tracking more effectively without introducing comparable costs.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Transformers face a topological limitation in dynamic state tracking because their feedforward architecture pushes evolving state representations deeper into layers until depth is exhausted, requiring a shift to recurrent architectures for implicit activation dynamics.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Transformers push evolving state representations deeper into their layers with each new input, exhausting depth and limiting dynamic tracking.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"511a5aa63d14ce0f754e2cf69a04d8df3f5e6aaff62e69690a068d474f1c5d87"},"source":{"id":"2604.17121","kind":"arxiv","version":3},"verdict":{"id":"ba4294d3-28b7-4f9a-bfcf-c24e21d7c5ec","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T06:31:05.070398Z","strongest_claim":"their purely feedforward architecture fundamentally limits dynamic state tracking. State tracking -- the iterative updating of latent variables reflecting an evolving environment -- involves inherently sequential dependencies that feedforward networks struggle to maintain. Consequently, feedforward models push evolving state representations deeper into their layer stack with each new input step, rendering information inaccessible in shallow layers and ultimately exhausting the model's depth.","one_line_summary":"Transformers face a topological limitation in dynamic state tracking because their feedforward architecture pushes evolving state representations deeper into layers until depth is exhausted, requiring a shift to recurrent architectures for implicit activation dynamics.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That dynamic depth models, explicit thinking traces, and latent thinking are inherently too computationally and memory inefficient to serve as scalable solutions, and that recurrent architectures will integrate state tracking more effectively without introducing comparable costs.","pith_extraction_headline":"Transformers push evolving state representations deeper into their layers with each new input, exhausting depth and limiting dynamic tracking."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.17121/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":"ba4294d3-28b7-4f9a-bfcf-c24e21d7c5ec"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-05T01:14:38Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"fWBBqm90cP0NXw+weIjJtFeriwdr6jzz6ZK5aRfSs4D4yqO3Kr8/pAG3sapJXfSnR7rGOTbPrzftvifcFQF2BA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-12T06:53:22.868273Z"},"content_sha256":"236b4f5e9a9aff5a82b8ca435de4d11cba71d107c390984e85e83816379b1f53","schema_version":"1.0","event_id":"sha256:236b4f5e9a9aff5a82b8ca435de4d11cba71d107c390984e85e83816379b1f53"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/BMM5EYN7NAZNSW7VKT2LCMP7T4/bundle.json","state_url":"https://pith.science/pith/BMM5EYN7NAZNSW7VKT2LCMP7T4/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/BMM5EYN7NAZNSW7VKT2LCMP7T4/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":"2026-06-12T06:53:22Z","links":{"resolver":"https://pith.science/pith/BMM5EYN7NAZNSW7VKT2LCMP7T4","bundle":"https://pith.science/pith/BMM5EYN7NAZNSW7VKT2LCMP7T4/bundle.json","state":"https://pith.science/pith/BMM5EYN7NAZNSW7VKT2LCMP7T4/state.json","well_known_bundle":"https://pith.science/.well-known/pith/BMM5EYN7NAZNSW7VKT2LCMP7T4/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:BMM5EYN7NAZNSW7VKT2LCMP7T4","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":"86cc7ddcbfb88ccb7429888589ecae5153cb07ffe5f9b45849786d86c5c71403","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-18T19:46:30Z","title_canon_sha256":"dc9ab62c6a6a71e3f67e75c33518c75c8f4a3620545822a512863e335212cefc"},"schema_version":"1.0","source":{"id":"2604.17121","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.17121","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"arxiv_version","alias_value":"2604.17121v3","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.17121","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_12","alias_value":"BMM5EYN7NAZN","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_16","alias_value":"BMM5EYN7NAZNSW7V","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_8","alias_value":"BMM5EYN7","created_at":"2026-06-05T01:14:38Z"}],"graph_snapshots":[{"event_id":"sha256:236b4f5e9a9aff5a82b8ca435de4d11cba71d107c390984e85e83816379b1f53","target":"graph","created_at":"2026-06-05T01:14:38Z","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":"their purely feedforward architecture fundamentally limits dynamic state tracking. State tracking -- the iterative updating of latent variables reflecting an evolving environment -- involves inherently sequential dependencies that feedforward networks struggle to maintain. Consequently, feedforward models push evolving state representations deeper into their layer stack with each new input step, rendering information inaccessible in shallow layers and ultimately exhausting the model's depth."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That dynamic depth models, explicit thinking traces, and latent thinking are inherently too computationally and memory inefficient to serve as scalable solutions, and that recurrent architectures will integrate state tracking more effectively without introducing comparable costs."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Transformers face a topological limitation in dynamic state tracking because their feedforward architecture pushes evolving state representations deeper into layers until depth is exhausted, requiring a shift to recurrent architectures for implicit activation dynamics."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Transformers push evolving state representations deeper into their layers with each new input, exhausting depth and limiting dynamic tracking."}],"snapshot_sha256":"511a5aa63d14ce0f754e2cf69a04d8df3f5e6aaff62e69690a068d474f1c5d87"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2604.17121/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Transformers encode structure in sequences via an expanding contextual history. However, their purely feedforward architecture fundamentally limits dynamic state tracking. State tracking -- the iterative updating of latent variables reflecting an evolving environment -- involves inherently sequential dependencies that feedforward networks struggle to maintain. Consequently, feedforward models push evolving state representations deeper into their layer stack with each new input step, rendering information inaccessible in shallow layers and ultimately exhausting the model's depth. While this dep","authors_text":"Michael C. Mozer, Rosanne Liu, Shoaib Ahmed Siddiqui","cross_cats":["cs.AI"],"headline":"Transformers push evolving state representations deeper into their layers with each new input, exhausting depth and limiting dynamic tracking.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-18T19:46:30Z","title":"The Topological Trouble With Transformers"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.17121","kind":"arxiv","version":3},"verdict":{"created_at":"2026-05-10T06:31:05.070398Z","id":"ba4294d3-28b7-4f9a-bfcf-c24e21d7c5ec","model_set":{"reader":"grok-4.3"},"one_line_summary":"Transformers face a topological limitation in dynamic state tracking because their feedforward architecture pushes evolving state representations deeper into layers until depth is exhausted, requiring a shift to recurrent architectures for implicit activation dynamics.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Transformers push evolving state representations deeper into their layers with each new input, exhausting depth and limiting dynamic tracking.","strongest_claim":"their purely feedforward architecture fundamentally limits dynamic state tracking. State tracking -- the iterative updating of latent variables reflecting an evolving environment -- involves inherently sequential dependencies that feedforward networks struggle to maintain. Consequently, feedforward models push evolving state representations deeper into their layer stack with each new input step, rendering information inaccessible in shallow layers and ultimately exhausting the model's depth.","weakest_assumption":"That dynamic depth models, explicit thinking traces, and latent thinking are inherently too computationally and memory inefficient to serve as scalable solutions, and that recurrent architectures will integrate state tracking more effectively without introducing comparable costs."}},"verdict_id":"ba4294d3-28b7-4f9a-bfcf-c24e21d7c5ec"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:413093d936351f110618857d09abc6a502783ffa173b02afec8b93cb319ff45b","target":"record","created_at":"2026-06-05T01:14:38Z","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":"86cc7ddcbfb88ccb7429888589ecae5153cb07ffe5f9b45849786d86c5c71403","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-18T19:46:30Z","title_canon_sha256":"dc9ab62c6a6a71e3f67e75c33518c75c8f4a3620545822a512863e335212cefc"},"schema_version":"1.0","source":{"id":"2604.17121","kind":"arxiv","version":3}},"canonical_sha256":"0b19d261bf6832d95bf554f4b131ff9f3902f55dbceb5e559f14ceb4abf67051","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0b19d261bf6832d95bf554f4b131ff9f3902f55dbceb5e559f14ceb4abf67051","first_computed_at":"2026-06-05T01:14:38.907997Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-05T01:14:38.907997Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"YHVGOotFijwRUVe6GZSmi6BC71/T5k+XHjvKYzMX/8P4pk337933dfXT61TlGttuppa8i53XmagzTHaIJmJ2Bg==","signature_status":"signed_v1","signed_at":"2026-06-05T01:14:38.908766Z","signed_message":"canonical_sha256_bytes"},"source_id":"2604.17121","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:413093d936351f110618857d09abc6a502783ffa173b02afec8b93cb319ff45b","sha256:236b4f5e9a9aff5a82b8ca435de4d11cba71d107c390984e85e83816379b1f53"],"state_sha256":"031804388fa843e0dd72eab26b9dcd77cec903a88be13f68d80615c81f7e689a"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"13frFJ6wp0WhOPBD1E767peI9HdueQa+T/Y77LAu9+aWuaEWziUKudi3vzVavD7mO5fFDCFGeSQ+gaO/WRgIDA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-12T06:53:22.870629Z","bundle_sha256":"13d3eed234a491a982372c8d3dedb482a20fbbc2cbb5aa0acea0c7153280934c"}}