{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:QIJUMINMVLUQZCVCUSBKH44OLO","short_pith_number":"pith:QIJUMINM","canonical_record":{"source":{"id":"2605.13772","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T16:48:48Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"d550e2ba716eb4453dcb5a3ef08eff2aca7e1dc83f0f84a33e5d634919dd7086","abstract_canon_sha256":"f0684a1fc6fe1216f578f79a2bc44445f0ca99fe984d73bf655dc9617e23c2e1"},"schema_version":"1.0"},"canonical_sha256":"82134621acaae90c8aa2a482a3f38e5bb70ec344c9d65f0533d90fe77489a6f4","source":{"kind":"arxiv","id":"2605.13772","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13772","created_at":"2026-05-18T02:44:15Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13772v1","created_at":"2026-05-18T02:44:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13772","created_at":"2026-05-18T02:44:15Z"},{"alias_kind":"pith_short_12","alias_value":"QIJUMINMVLUQ","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"QIJUMINMVLUQZCVC","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"QIJUMINM","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:QIJUMINMVLUQZCVCUSBKH44OLO","target":"record","payload":{"canonical_record":{"source":{"id":"2605.13772","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T16:48:48Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"d550e2ba716eb4453dcb5a3ef08eff2aca7e1dc83f0f84a33e5d634919dd7086","abstract_canon_sha256":"f0684a1fc6fe1216f578f79a2bc44445f0ca99fe984d73bf655dc9617e23c2e1"},"schema_version":"1.0"},"canonical_sha256":"82134621acaae90c8aa2a482a3f38e5bb70ec344c9d65f0533d90fe77489a6f4","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:44:15.981423Z","signature_b64":"LXbcnPnYh3cpNJCb/uxIwFGOMUWbe7yQU1CLyi97UaY3GliFeGbthMhznXSQ9x/Lws3PBFnO378WhpyB6HaGAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"82134621acaae90c8aa2a482a3f38e5bb70ec344c9d65f0533d90fe77489a6f4","last_reissued_at":"2026-05-18T02:44:15.980810Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:44:15.980810Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.13772","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Uirmx5jamd2OsFTeQ05kYNRK0M97uQoxnnNbcJJ1kB1gw+5p5ZeZXDWYflYHyO3WF8/0L/0VC4p900TgSDzHAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-07T12:39:59.730491Z"},"content_sha256":"63e1e0a1d303eb0f7ed2010103d09b4f4d48b6b12a3722012f872e763360a3ee","schema_version":"1.0","event_id":"sha256:63e1e0a1d303eb0f7ed2010103d09b4f4d48b6b12a3722012f872e763360a3ee"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:QIJUMINMVLUQZCVCUSBKH44OLO","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Where Does Reasoning Break? Step-Level Hallucination Detection via Hidden-State Transport Geometry","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Hidden-state trajectories expose the exact step where LLM reasoning first breaks via transport geometry.","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Ali Baheri, Tyler Alvarez","submitted_at":"2026-05-13T16:48:48Z","abstract_excerpt":"Large language models hallucinate during multi-step reasoning, but most existing detectors operate at the trace level: they assign one confidence score to a full output, fail to localize the first error, and often require multiple sampled completions. We frame hallucination instead as a property of the hidden-state trajectory produced during a single forward pass. Correct reasoning moves through a stable manifold of locally coherent transitions; a first error appears as a localized excursion in transport cost away from this manifold. We operationalize this view with a label-conditioned teacher"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We prove that contrastive PCA is the optimal projection for a transport-separation objective between first error and correct states, and that single-pass first error localization holds whenever the first error creates a positive transport margin over preceding correct transitions.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that the first error creates a positive transport margin over preceding correct transitions (stated in the abstract) and that a stable manifold of locally coherent transitions exists for correct reasoning.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Hallucination is detected as a transport-cost excursion in hidden-state trajectories, localized via contrastive PCA in a teacher model and distilled to a BiLSTM student.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Hidden-state trajectories expose the exact step where LLM reasoning first breaks via transport geometry.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"efa85fefcb1c98e682359bfe75e56e9bc0a0f08ed1b87ed356da560e42600c2f"},"source":{"id":"2605.13772","kind":"arxiv","version":1},"verdict":{"id":"e4ea1508-098a-4573-822e-54e52dc27c96","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:11:26.310867Z","strongest_claim":"We prove that contrastive PCA is the optimal projection for a transport-separation objective between first error and correct states, and that single-pass first error localization holds whenever the first error creates a positive transport margin over preceding correct transitions.","one_line_summary":"Hallucination is detected as a transport-cost excursion in hidden-state trajectories, localized via contrastive PCA in a teacher model and distilled to a BiLSTM student.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that the first error creates a positive transport margin over preceding correct transitions (stated in the abstract) and that a stable manifold of locally coherent transitions exists for correct reasoning.","pith_extraction_headline":"Hidden-state trajectories expose the exact step where LLM reasoning first breaks via transport geometry."},"references":{"count":27,"sample":[{"doi":"","year":2018,"title":"Abid, A., Zhang, M. J., Bagaria, V . K., and Zou, J. Exploring patterns enriched in a dataset with contrastive principal component analysis.Nature Communications, 9(1):2134, 2018. doi: 10.1038/ s41467","work_id":"08a3108f-0520-4001-82e3-74d6f866aa0a","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.48550/arxiv.2602.16015","year":2026,"title":"Geometry-aware uncertainty quantification via conformal prediction on man- ifolds","work_id":"77604c07-4e88-4584-8c33-e349b35787b7","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Azaria, A. and Mitchell, T. The internal state of an LLM knows when it’s lying. InFindings of the Asso- ciation for Computational Linguistics: EMNLP 2023, pp. 967–976, Singapore, 2023","work_id":"d8c63f53-f165-4631-a2d0-64fb0195b233","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.48550/arxiv.2602","year":2026,"title":"Merlean: An agentic framework for autoformalization in quantum computation","work_id":"2c62ea97-8f54-46ed-ab36-91703f30ac87","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Baheri, A. and Alm, C. O. LLMs-augmented contex- tual bandit. InNeurIPS 2023 Workshop on Foundation Models for Decision Making, 2023. FMDM@NeurIPS 2023","work_id":"c7296444-dffa-43dc-8460-5108a6ecd4ff","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":27,"snapshot_sha256":"e6d4ed96f98c7adfb813d40d01a1d4605af1d3bd8f6408fc203703241b15df0d","internal_anchors":5},"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":"e4ea1508-098a-4573-822e-54e52dc27c96"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+nOqCXerBy+i57Z91eL/1LyUeNUtoKs80axZRqGlp2e6LAJ23CCbhIC2YUEJPWi/lHcxXAPTFAj/yLBesb7HAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-07T12:39:59.731367Z"},"content_sha256":"4aab7ead6c88f6509cbe0add1affbee1197d1edab4120f25dd0e7d4085c4b00c","schema_version":"1.0","event_id":"sha256:4aab7ead6c88f6509cbe0add1affbee1197d1edab4120f25dd0e7d4085c4b00c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/QIJUMINMVLUQZCVCUSBKH44OLO/bundle.json","state_url":"https://pith.science/pith/QIJUMINMVLUQZCVCUSBKH44OLO/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/QIJUMINMVLUQZCVCUSBKH44OLO/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":"2026-06-07T12:39:59Z","links":{"resolver":"https://pith.science/pith/QIJUMINMVLUQZCVCUSBKH44OLO","bundle":"https://pith.science/pith/QIJUMINMVLUQZCVCUSBKH44OLO/bundle.json","state":"https://pith.science/pith/QIJUMINMVLUQZCVCUSBKH44OLO/state.json","well_known_bundle":"https://pith.science/.well-known/pith/QIJUMINMVLUQZCVCUSBKH44OLO/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:QIJUMINMVLUQZCVCUSBKH44OLO","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":"f0684a1fc6fe1216f578f79a2bc44445f0ca99fe984d73bf655dc9617e23c2e1","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T16:48:48Z","title_canon_sha256":"d550e2ba716eb4453dcb5a3ef08eff2aca7e1dc83f0f84a33e5d634919dd7086"},"schema_version":"1.0","source":{"id":"2605.13772","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13772","created_at":"2026-05-18T02:44:15Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13772v1","created_at":"2026-05-18T02:44:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13772","created_at":"2026-05-18T02:44:15Z"},{"alias_kind":"pith_short_12","alias_value":"QIJUMINMVLUQ","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"QIJUMINMVLUQZCVC","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"QIJUMINM","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:4aab7ead6c88f6509cbe0add1affbee1197d1edab4120f25dd0e7d4085c4b00c","target":"graph","created_at":"2026-05-18T02:44:15Z","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":"We prove that contrastive PCA is the optimal projection for a transport-separation objective between first error and correct states, and that single-pass first error localization holds whenever the first error creates a positive transport margin over preceding correct transitions."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The assumption that the first error creates a positive transport margin over preceding correct transitions (stated in the abstract) and that a stable manifold of locally coherent transitions exists for correct reasoning."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Hallucination is detected as a transport-cost excursion in hidden-state trajectories, localized via contrastive PCA in a teacher model and distilled to a BiLSTM student."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Hidden-state trajectories expose the exact step where LLM reasoning first breaks via transport geometry."}],"snapshot_sha256":"efa85fefcb1c98e682359bfe75e56e9bc0a0f08ed1b87ed356da560e42600c2f"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Large language models hallucinate during multi-step reasoning, but most existing detectors operate at the trace level: they assign one confidence score to a full output, fail to localize the first error, and often require multiple sampled completions. We frame hallucination instead as a property of the hidden-state trajectory produced during a single forward pass. Correct reasoning moves through a stable manifold of locally coherent transitions; a first error appears as a localized excursion in transport cost away from this manifold. We operationalize this view with a label-conditioned teacher","authors_text":"Ali Baheri, Tyler Alvarez","cross_cats":["cs.AI"],"headline":"Hidden-state trajectories expose the exact step where LLM reasoning first breaks via transport geometry.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T16:48:48Z","title":"Where Does Reasoning Break? Step-Level Hallucination Detection via Hidden-State Transport Geometry"},"references":{"count":27,"internal_anchors":5,"resolved_work":27,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Abid, A., Zhang, M. J., Bagaria, V . K., and Zou, J. Exploring patterns enriched in a dataset with contrastive principal component analysis.Nature Communications, 9(1):2134, 2018. doi: 10.1038/ s41467","work_id":"08a3108f-0520-4001-82e3-74d6f866aa0a","year":2018},{"cited_arxiv_id":"","doi":"10.48550/arxiv.2602.16015","is_internal_anchor":false,"ref_index":2,"title":"Geometry-aware uncertainty quantification via conformal prediction on man- ifolds","work_id":"77604c07-4e88-4584-8c33-e349b35787b7","year":2026},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Azaria, A. and Mitchell, T. The internal state of an LLM knows when it’s lying. InFindings of the Asso- ciation for Computational Linguistics: EMNLP 2023, pp. 967–976, Singapore, 2023","work_id":"d8c63f53-f165-4631-a2d0-64fb0195b233","year":2023},{"cited_arxiv_id":"","doi":"10.48550/arxiv.2602","is_internal_anchor":false,"ref_index":4,"title":"Merlean: An agentic framework for autoformalization in quantum computation","work_id":"2c62ea97-8f54-46ed-ab36-91703f30ac87","year":2026},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Baheri, A. and Alm, C. O. LLMs-augmented contex- tual bandit. InNeurIPS 2023 Workshop on Foundation Models for Decision Making, 2023. FMDM@NeurIPS 2023","work_id":"c7296444-dffa-43dc-8460-5108a6ecd4ff","year":2023}],"snapshot_sha256":"e6d4ed96f98c7adfb813d40d01a1d4605af1d3bd8f6408fc203703241b15df0d"},"source":{"id":"2605.13772","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-14T19:11:26.310867Z","id":"e4ea1508-098a-4573-822e-54e52dc27c96","model_set":{"reader":"grok-4.3"},"one_line_summary":"Hallucination is detected as a transport-cost excursion in hidden-state trajectories, localized via contrastive PCA in a teacher model and distilled to a BiLSTM student.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Hidden-state trajectories expose the exact step where LLM reasoning first breaks via transport geometry.","strongest_claim":"We prove that contrastive PCA is the optimal projection for a transport-separation objective between first error and correct states, and that single-pass first error localization holds whenever the first error creates a positive transport margin over preceding correct transitions.","weakest_assumption":"The assumption that the first error creates a positive transport margin over preceding correct transitions (stated in the abstract) and that a stable manifold of locally coherent transitions exists for correct reasoning."}},"verdict_id":"e4ea1508-098a-4573-822e-54e52dc27c96"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:63e1e0a1d303eb0f7ed2010103d09b4f4d48b6b12a3722012f872e763360a3ee","target":"record","created_at":"2026-05-18T02:44:15Z","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":"f0684a1fc6fe1216f578f79a2bc44445f0ca99fe984d73bf655dc9617e23c2e1","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T16:48:48Z","title_canon_sha256":"d550e2ba716eb4453dcb5a3ef08eff2aca7e1dc83f0f84a33e5d634919dd7086"},"schema_version":"1.0","source":{"id":"2605.13772","kind":"arxiv","version":1}},"canonical_sha256":"82134621acaae90c8aa2a482a3f38e5bb70ec344c9d65f0533d90fe77489a6f4","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"82134621acaae90c8aa2a482a3f38e5bb70ec344c9d65f0533d90fe77489a6f4","first_computed_at":"2026-05-18T02:44:15.980810Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T02:44:15.980810Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"LXbcnPnYh3cpNJCb/uxIwFGOMUWbe7yQU1CLyi97UaY3GliFeGbthMhznXSQ9x/Lws3PBFnO378WhpyB6HaGAQ==","signature_status":"signed_v1","signed_at":"2026-05-18T02:44:15.981423Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.13772","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:63e1e0a1d303eb0f7ed2010103d09b4f4d48b6b12a3722012f872e763360a3ee","sha256:4aab7ead6c88f6509cbe0add1affbee1197d1edab4120f25dd0e7d4085c4b00c"],"state_sha256":"817c2d4028cadae732b8c3431e659cb215b948504ac31345246d7ddaff4719bb"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"kt2H5zLjM0xrMS8xytxzI87kwS13oU9BaZXrUDIhG7MlP0WQdWJIOFmB0LOcYmYyE+bf3cM/rJm+KSKo2TqgCg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-07T12:39:59.735119Z","bundle_sha256":"a18f99a8e6089aef18c3e26ea9af745662499c96e51bdd69399a66a4a80dd91e"}}