{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QNGEPJOCAWSYZHSBTWBC7BYFT4","short_pith_number":"pith:QNGEPJOC","schema_version":"1.0","canonical_sha256":"834c47a5c205a58c9e419d822f87059f1d0f9e31aec1e38865b18bc70ae16ad3","source":{"kind":"arxiv","id":"2608.10145","version":1},"attestation_state":"computed","paper":{"title":"The Evaluation Protocol Determines the Result: An Independent Reproduction of LeWorldModel on TwoRoom","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Joyjeet Singh","submitted_at":"2026-08-10T19:00:51Z","abstract_excerpt":"LeWorldModel trains a latent world model with a prediction loss and a single anti-collapse regulariser, and reports approximately 87% of goals reached on TwoRoom, its simplest diagnostic environment. We reproduce that result by independent reimplementation on roughly $25 of rented compute, with all evaluation on one laptop CPU.\n  We reach 94.0% at the repository's evaluation goal offset, against 84.0% for the authors' own released checkpoint measured under our protocol on identical episodes, and we reproduce the reported representation result directly (position probe Pearson r = 0.9988 against"},"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":"2608.10145","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-08-10T19:00:51Z","cross_cats_sorted":[],"title_canon_sha256":"fa31f438865bb5bf9ab9cd9a480fa46ec4c72915c5975c897efa37422a6be7ca","abstract_canon_sha256":"7ceaf10363196671c56c1ade27e0092944adb8f9efefcfcdbb7ea5120f67a1f3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-12T00:22:54.371363Z","signature_b64":"SSlUlO6a6nszHXgloxe4pJXYfPTGVvYCSm0I9DT4hZeg5SUe21gZfdk8ZaksFDZ48H3clU8AjS7I/YGqtkeDAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"834c47a5c205a58c9e419d822f87059f1d0f9e31aec1e38865b18bc70ae16ad3","last_reissued_at":"2026-08-12T00:22:54.369205Z","signature_status":"signed_v1","first_computed_at":"2026-08-12T00:22:54.369205Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Evaluation Protocol Determines the Result: An Independent Reproduction of LeWorldModel on TwoRoom","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Joyjeet Singh","submitted_at":"2026-08-10T19:00:51Z","abstract_excerpt":"LeWorldModel trains a latent world model with a prediction loss and a single anti-collapse regulariser, and reports approximately 87% of goals reached on TwoRoom, its simplest diagnostic environment. We reproduce that result by independent reimplementation on roughly $25 of rented compute, with all evaluation on one laptop CPU.\n  We reach 94.0% at the repository's evaluation goal offset, against 84.0% for the authors' own released checkpoint measured under our protocol on identical episodes, and we reproduce the reported representation result directly (position probe Pearson r = 0.9988 against"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.10145","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2608.10145/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2608.10145","created_at":"2026-08-12T00:22:54.372309+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.10145v1","created_at":"2026-08-12T00:22:54.372309+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.10145","created_at":"2026-08-12T00:22:54.372309+00:00"},{"alias_kind":"pith_short_12","alias_value":"QNGEPJOCAWSY","created_at":"2026-08-12T00:22:54.372309+00:00"},{"alias_kind":"pith_short_16","alias_value":"QNGEPJOCAWSYZHSB","created_at":"2026-08-12T00:22:54.372309+00:00"},{"alias_kind":"pith_short_8","alias_value":"QNGEPJOC","created_at":"2026-08-12T00:22:54.372309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2608.12959","citing_title":"The Objective Is the Bottleneck: Latent World Models Encode What Their Planners Cannot Use","ref_index":3,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QNGEPJOCAWSYZHSBTWBC7BYFT4","json":"https://pith.science/pith/QNGEPJOCAWSYZHSBTWBC7BYFT4.json","graph_json":"https://pith.science/api/pith-number/QNGEPJOCAWSYZHSBTWBC7BYFT4/graph.json","events_json":"https://pith.science/api/pith-number/QNGEPJOCAWSYZHSBTWBC7BYFT4/events.json","paper":"https://pith.science/paper/QNGEPJOC"},"agent_actions":{"view_html":"https://pith.science/pith/QNGEPJOCAWSYZHSBTWBC7BYFT4","download_json":"https://pith.science/pith/QNGEPJOCAWSYZHSBTWBC7BYFT4.json","view_paper":"https://pith.science/paper/QNGEPJOC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.10145&json=true","fetch_graph":"https://pith.science/api/pith-number/QNGEPJOCAWSYZHSBTWBC7BYFT4/graph.json","fetch_events":"https://pith.science/api/pith-number/QNGEPJOCAWSYZHSBTWBC7BYFT4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QNGEPJOCAWSYZHSBTWBC7BYFT4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QNGEPJOCAWSYZHSBTWBC7BYFT4/action/storage_attestation","attest_author":"https://pith.science/pith/QNGEPJOCAWSYZHSBTWBC7BYFT4/action/author_attestation","sign_citation":"https://pith.science/pith/QNGEPJOCAWSYZHSBTWBC7BYFT4/action/citation_signature","submit_replication":"https://pith.science/pith/QNGEPJOCAWSYZHSBTWBC7BYFT4/action/replication_record"}},"created_at":"2026-08-12T00:22:54.372309+00:00","updated_at":"2026-08-12T00:22:54.372309+00:00"}