{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:PAK2NUIABZI2AWNA2VWHUGYKUL","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":"6987374dc451519ff851919f375b21ea8a49bf936b0529d23c427e96cd1aa356","cross_cats_sorted":["cs.LG","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-14T22:37:33Z","title_canon_sha256":"2ee461029fdeeeebb789b1d76858956f16d334d95a202ce589669959549b5ff3"},"schema_version":"1.0","source":{"id":"2605.15454","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.15454","created_at":"2026-05-20T00:00:59Z"},{"alias_kind":"arxiv_version","alias_value":"2605.15454v1","created_at":"2026-05-20T00:00:59Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15454","created_at":"2026-05-20T00:00:59Z"},{"alias_kind":"pith_short_12","alias_value":"PAK2NUIABZI2","created_at":"2026-05-20T00:00:59Z"},{"alias_kind":"pith_short_16","alias_value":"PAK2NUIABZI2AWNA","created_at":"2026-05-20T00:00:59Z"},{"alias_kind":"pith_short_8","alias_value":"PAK2NUIA","created_at":"2026-05-20T00:00:59Z"}],"graph_snapshots":[{"event_id":"sha256:e7fa5355dc0b1c4260b2b0a96a717dec30b8e4f635430e98590c2a9af40884b1","target":"graph","created_at":"2026-05-20T00:00:59Z","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":"After residualizing trajectory statistics on length, difficulty remains systematically coupled to corrected trajectory geometry across all domains studied. The clearest reasoning-specific separation appears in the code domain, where harder problems show more direct corrected trajectories and less heterogeneous local curvature in reasoning-trained models than in matched instruction-tuned baselines."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That residualizing raw trajectory statistics on generation length successfully removes all mechanical length-induced effects and leaves only the component attributable to reasoning strategy or internal computation."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"After length-correcting hidden-state trajectories during chain-of-thought, reasoning models show systematically different geometry on harder problems than baselines, strongest in competitive programming."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"After length correction, reasoning models show distinct hidden-state trajectories on harder problems."}],"snapshot_sha256":"85f9268695eb2cfd6131799219a41f072b599757cd8b2ec8981df2744cae394f"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"a92230f44cfaf3190036fd61fcc4d094ec508df946b5ca722110c1747436e564"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"cited_work_retraction","ran_at":"2026-05-19T15:51:55.289286Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"citation_quote_validity","ran_at":"2026-05-19T15:49:52.204121Z","status":"completed","version":"0.1.0"},{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T15:01:17.612094Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T14:50:17.313748Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T14:21:54.106172Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T13:33:22.675294Z","status":"skipped","version":"1.0.0"}],"endpoint":"/pith/2605.15454/integrity.json","findings":[],"snapshot_sha256":"0f17cfab8e6dac1d12ecc3804fe9e585399e81eacf7ed7ab52ad8cdf7f746b79","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reasoning-trained language models often spend more tokens on harder problems, but longer chains of thought do not show whether a model is merely computing for more steps or following a different internal trajectory. We study this distinction through hidden-state trajectories during chain-of-thought generation across competitive programming, mathematics, and Boolean satisfiability. Raw trajectory geometry is strongly shaped by generation length: longer generations mechanically alter path statistics, so difficulty-dependent comparisons are misleading without adjustment. After residualizing traje","authors_text":"Anders Gj{\\o}lbye, Lars Kai Hansen, Sanmi Koyejo","cross_cats":["cs.LG","stat.ML"],"headline":"After length correction, reasoning models show distinct hidden-state trajectories on harder problems.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-14T22:37:33Z","title":"Reasoning Models Don't Just Think Longer, They Move Differently"},"references":{"count":14,"internal_anchors":2,"resolved_work":14,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Mitigating overthinking in large reasoning models via manifold steering","work_id":"f8c2a7ce-0438-463d-8985-cfd021da58cb","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Jolliffe and Jorge Cadima","work_id":"5b863ffc-2076-49ed-889b-81ecc3cf2f65","year":2065},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Probing the difficulty perception mechanism of large language models","work_id":"b131cbd5-0794-4a78-a1db-5d7f274ea2fb","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"LLMs encode how difficult problems are.arXiv preprint arXiv:2510.18147,","work_id":"6948d73b-e278-4096-8f01-0856a18ca1ba","year":null},{"cited_arxiv_id":"2408.03314","doi":"","is_internal_anchor":true,"ref_index":5,"title":"Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters","work_id":"a8d50b24-bdf5-46ed-bc4f-2927dfd81f1d","year":null}],"snapshot_sha256":"d8e12ec78efbfd91823462f19e54c49c88ddef77fd875e4bb3bf612409b65a30"},"source":{"id":"2605.15454","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T14:39:02.963239Z","id":"7b502532-69c7-4d89-bdda-a77467dddae6","model_set":{"reader":"grok-4.3"},"one_line_summary":"After length-correcting hidden-state trajectories during chain-of-thought, reasoning models show systematically different geometry on harder problems than baselines, strongest in competitive programming.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"After length correction, reasoning models show distinct hidden-state trajectories on harder problems.","strongest_claim":"After residualizing trajectory statistics on length, difficulty remains systematically coupled to corrected trajectory geometry across all domains studied. The clearest reasoning-specific separation appears in the code domain, where harder problems show more direct corrected trajectories and less heterogeneous local curvature in reasoning-trained models than in matched instruction-tuned baselines.","weakest_assumption":"That residualizing raw trajectory statistics on generation length successfully removes all mechanical length-induced effects and leaves only the component attributable to reasoning strategy or internal computation."}},"verdict_id":"7b502532-69c7-4d89-bdda-a77467dddae6"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:483e843bd8465c62aa5d661fb8da59e81bad8d84aec1e8429ce34ccc73f1e841","target":"record","created_at":"2026-05-20T00:00:59Z","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":"6987374dc451519ff851919f375b21ea8a49bf936b0529d23c427e96cd1aa356","cross_cats_sorted":["cs.LG","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-14T22:37:33Z","title_canon_sha256":"2ee461029fdeeeebb789b1d76858956f16d334d95a202ce589669959549b5ff3"},"schema_version":"1.0","source":{"id":"2605.15454","kind":"arxiv","version":1}},"canonical_sha256":"7815a6d1000e51a059a0d56c7a1b0aa2ea64bc9151b67e3e64bd2f69f0deedcc","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7815a6d1000e51a059a0d56c7a1b0aa2ea64bc9151b67e3e64bd2f69f0deedcc","first_computed_at":"2026-05-20T00:00:59.437691Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:00:59.437691Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"nMElWnR+8caqGNuicC2TF+S9CdGY+E3Cz3ipxSLG4ar+xD7udMKKriB31R/EwqY2XmYgSBpR+GEqnZwZlo5jAw==","signature_status":"signed_v1","signed_at":"2026-05-20T00:00:59.438454Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.15454","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:483e843bd8465c62aa5d661fb8da59e81bad8d84aec1e8429ce34ccc73f1e841","sha256:e7fa5355dc0b1c4260b2b0a96a717dec30b8e4f635430e98590c2a9af40884b1"],"state_sha256":"0153c8f5187bfc23e942ae9c59e2f0625533225fb5835452fbbaa4bbc016435a"}