{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:GYDPBSIQ5W7FS3RIBYP677VZFS","short_pith_number":"pith:GYDPBSIQ","canonical_record":{"source":{"id":"2607.01033","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-01T15:01:30Z","cross_cats_sorted":[],"title_canon_sha256":"0ac165cbcf690c99f4ef863a6e2211b5f73fd510f340d0a8ca94f948959107dd","abstract_canon_sha256":"1815f3f05638a0c4aff7a0a69e5558d195546d03b3dc923ffcd10b503aedfa99"},"schema_version":"1.0"},"canonical_sha256":"3606f0c910edbe596e280e1feffeb92c8f3740ae703c29c111d7fb1df00c2b30","source":{"kind":"arxiv","id":"2607.01033","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.01033","created_at":"2026-07-02T01:18:26Z"},{"alias_kind":"arxiv_version","alias_value":"2607.01033v1","created_at":"2026-07-02T01:18:26Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.01033","created_at":"2026-07-02T01:18:26Z"},{"alias_kind":"pith_short_12","alias_value":"GYDPBSIQ5W7F","created_at":"2026-07-02T01:18:26Z"},{"alias_kind":"pith_short_16","alias_value":"GYDPBSIQ5W7FS3RI","created_at":"2026-07-02T01:18:26Z"},{"alias_kind":"pith_short_8","alias_value":"GYDPBSIQ","created_at":"2026-07-02T01:18:26Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:GYDPBSIQ5W7FS3RIBYP677VZFS","target":"record","payload":{"canonical_record":{"source":{"id":"2607.01033","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-01T15:01:30Z","cross_cats_sorted":[],"title_canon_sha256":"0ac165cbcf690c99f4ef863a6e2211b5f73fd510f340d0a8ca94f948959107dd","abstract_canon_sha256":"1815f3f05638a0c4aff7a0a69e5558d195546d03b3dc923ffcd10b503aedfa99"},"schema_version":"1.0"},"canonical_sha256":"3606f0c910edbe596e280e1feffeb92c8f3740ae703c29c111d7fb1df00c2b30","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-02T01:18:26.594441Z","signature_b64":"suVCz4wr9wA7nUgDyzEgVL/3aLdrttACPGN9y7H3yiWJ2j2Vb/4zb7IG69skSGfbOHIKu3rD1uMabCy7CcdvCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3606f0c910edbe596e280e1feffeb92c8f3740ae703c29c111d7fb1df00c2b30","last_reissued_at":"2026-07-02T01:18:26.594021Z","signature_status":"signed_v1","first_computed_at":"2026-07-02T01:18:26.594021Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2607.01033","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":"2026-07-02T01:18:26Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"OfSXfhKA01nYkTz2zL06oltekm9ZmuvnJytiEzbLMNQQl+EblLXmqny2zREKNtljWempYeFt9u0FMRf+FH3RCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-03T11:21:51.628899Z"},"content_sha256":"8eaba6e03aa9e1edfd5c1ee0ab8d323d45baf3d84b27f3d683166537eb4438cf","schema_version":"1.0","event_id":"sha256:8eaba6e03aa9e1edfd5c1ee0ab8d323d45baf3d84b27f3d683166537eb4438cf"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:GYDPBSIQ5W7FS3RIBYP677VZFS","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"The Model Organism Lottery: Model Organism Interpretability Strongly Depends on Training Methodology","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Andrzej Szablewski, Gabriel Konar-Steenberg, Nikita Menon, Raffaello Fornasiere, Stefan Heimersheim","submitted_at":"2026-07-01T15:01:30Z","abstract_excerpt":"Model organisms (MOs) - language models trained to exhibit undesired or unnatural behaviours - are frequently used as testbeds for evaluating white-box interpretability techniques. Current MOs are typically constructed via post-hoc supervised fine-tuning (SFT) on behavioural transcripts or synthetic documents. Prior research has shown that interpretability methods can easily identify hidden behaviours in these MOs. However, recent work suggests that such post-hoc training methods may make interpretability unrealistically easy. We investigate this claim by constructing a suite of 54 $\\verb|OLMo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.01033","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/2607.01033/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":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-02T01:18:26Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rmH/6X6AipnhkYWCypcp/VvrwJ/b88s9Cm21Pw9u1TDeNkRrqbpbxFvM78p/Jje/pOLgdLZ98HT6vCLsQiY0Bg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-03T11:21:51.629268Z"},"content_sha256":"7f0f7e4d619da941004959b40ef87d4f9f9596696bf5c71eb87e24489aae783c","schema_version":"1.0","event_id":"sha256:7f0f7e4d619da941004959b40ef87d4f9f9596696bf5c71eb87e24489aae783c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/GYDPBSIQ5W7FS3RIBYP677VZFS/bundle.json","state_url":"https://pith.science/pith/GYDPBSIQ5W7FS3RIBYP677VZFS/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/GYDPBSIQ5W7FS3RIBYP677VZFS/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":"2026-07-03T11:21:51Z","links":{"resolver":"https://pith.science/pith/GYDPBSIQ5W7FS3RIBYP677VZFS","bundle":"https://pith.science/pith/GYDPBSIQ5W7FS3RIBYP677VZFS/bundle.json","state":"https://pith.science/pith/GYDPBSIQ5W7FS3RIBYP677VZFS/state.json","well_known_bundle":"https://pith.science/.well-known/pith/GYDPBSIQ5W7FS3RIBYP677VZFS/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:GYDPBSIQ5W7FS3RIBYP677VZFS","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":"1815f3f05638a0c4aff7a0a69e5558d195546d03b3dc923ffcd10b503aedfa99","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-01T15:01:30Z","title_canon_sha256":"0ac165cbcf690c99f4ef863a6e2211b5f73fd510f340d0a8ca94f948959107dd"},"schema_version":"1.0","source":{"id":"2607.01033","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.01033","created_at":"2026-07-02T01:18:26Z"},{"alias_kind":"arxiv_version","alias_value":"2607.01033v1","created_at":"2026-07-02T01:18:26Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.01033","created_at":"2026-07-02T01:18:26Z"},{"alias_kind":"pith_short_12","alias_value":"GYDPBSIQ5W7F","created_at":"2026-07-02T01:18:26Z"},{"alias_kind":"pith_short_16","alias_value":"GYDPBSIQ5W7FS3RI","created_at":"2026-07-02T01:18:26Z"},{"alias_kind":"pith_short_8","alias_value":"GYDPBSIQ","created_at":"2026-07-02T01:18:26Z"}],"graph_snapshots":[{"event_id":"sha256:7f0f7e4d619da941004959b40ef87d4f9f9596696bf5c71eb87e24489aae783c","target":"graph","created_at":"2026-07-02T01:18:26Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2607.01033/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Model organisms (MOs) - language models trained to exhibit undesired or unnatural behaviours - are frequently used as testbeds for evaluating white-box interpretability techniques. Current MOs are typically constructed via post-hoc supervised fine-tuning (SFT) on behavioural transcripts or synthetic documents. Prior research has shown that interpretability methods can easily identify hidden behaviours in these MOs. However, recent work suggests that such post-hoc training methods may make interpretability unrealistically easy. We investigate this claim by constructing a suite of 54 $\\verb|OLMo","authors_text":"Andrzej Szablewski, Gabriel Konar-Steenberg, Nikita Menon, Raffaello Fornasiere, Stefan Heimersheim","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-01T15:01:30Z","title":"The Model Organism Lottery: Model Organism Interpretability Strongly Depends on Training Methodology"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.01033","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:8eaba6e03aa9e1edfd5c1ee0ab8d323d45baf3d84b27f3d683166537eb4438cf","target":"record","created_at":"2026-07-02T01:18:26Z","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":"1815f3f05638a0c4aff7a0a69e5558d195546d03b3dc923ffcd10b503aedfa99","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-01T15:01:30Z","title_canon_sha256":"0ac165cbcf690c99f4ef863a6e2211b5f73fd510f340d0a8ca94f948959107dd"},"schema_version":"1.0","source":{"id":"2607.01033","kind":"arxiv","version":1}},"canonical_sha256":"3606f0c910edbe596e280e1feffeb92c8f3740ae703c29c111d7fb1df00c2b30","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3606f0c910edbe596e280e1feffeb92c8f3740ae703c29c111d7fb1df00c2b30","first_computed_at":"2026-07-02T01:18:26.594021Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-02T01:18:26.594021Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"suVCz4wr9wA7nUgDyzEgVL/3aLdrttACPGN9y7H3yiWJ2j2Vb/4zb7IG69skSGfbOHIKu3rD1uMabCy7CcdvCA==","signature_status":"signed_v1","signed_at":"2026-07-02T01:18:26.594441Z","signed_message":"canonical_sha256_bytes"},"source_id":"2607.01033","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:8eaba6e03aa9e1edfd5c1ee0ab8d323d45baf3d84b27f3d683166537eb4438cf","sha256:7f0f7e4d619da941004959b40ef87d4f9f9596696bf5c71eb87e24489aae783c"],"state_sha256":"0f72049c5e909e84bb88c76cbb746bf65119b17e88c5b96b7d3ad651fb2ac0bb"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ePwK72gH1q6KZdn2vEurqNb0YYWWDZdaa+9AU62Zk9XPEnTN99MP9buIfSvzVaRpMRRljtN9p9BAYsHndW+fDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-03T11:21:51.631162Z","bundle_sha256":"fd49cc0412a10b6ef835f3f4177c6a5a0344af08dd0ca5e1b0b0ff15d2a10fd5"}}