{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:GYDPBSIQ5W7FS3RIBYP677VZFS","short_pith_number":"pith:GYDPBSIQ","schema_version":"1.0","canonical_sha256":"3606f0c910edbe596e280e1feffeb92c8f3740ae703c29c111d7fb1df00c2b30","source":{"kind":"arxiv","id":"2607.01033","version":1},"attestation_state":"computed","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"},"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":"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"},"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"},"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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2607.01033","created_at":"2026-07-02T01:18:26.594087+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.01033v1","created_at":"2026-07-02T01:18:26.594087+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.01033","created_at":"2026-07-02T01:18:26.594087+00:00"},{"alias_kind":"pith_short_12","alias_value":"GYDPBSIQ5W7F","created_at":"2026-07-02T01:18:26.594087+00:00"},{"alias_kind":"pith_short_16","alias_value":"GYDPBSIQ5W7FS3RI","created_at":"2026-07-02T01:18:26.594087+00:00"},{"alias_kind":"pith_short_8","alias_value":"GYDPBSIQ","created_at":"2026-07-02T01:18:26.594087+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GYDPBSIQ5W7FS3RIBYP677VZFS","json":"https://pith.science/pith/GYDPBSIQ5W7FS3RIBYP677VZFS.json","graph_json":"https://pith.science/api/pith-number/GYDPBSIQ5W7FS3RIBYP677VZFS/graph.json","events_json":"https://pith.science/api/pith-number/GYDPBSIQ5W7FS3RIBYP677VZFS/events.json","paper":"https://pith.science/paper/GYDPBSIQ"},"agent_actions":{"view_html":"https://pith.science/pith/GYDPBSIQ5W7FS3RIBYP677VZFS","download_json":"https://pith.science/pith/GYDPBSIQ5W7FS3RIBYP677VZFS.json","view_paper":"https://pith.science/paper/GYDPBSIQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.01033&json=true","fetch_graph":"https://pith.science/api/pith-number/GYDPBSIQ5W7FS3RIBYP677VZFS/graph.json","fetch_events":"https://pith.science/api/pith-number/GYDPBSIQ5W7FS3RIBYP677VZFS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GYDPBSIQ5W7FS3RIBYP677VZFS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GYDPBSIQ5W7FS3RIBYP677VZFS/action/storage_attestation","attest_author":"https://pith.science/pith/GYDPBSIQ5W7FS3RIBYP677VZFS/action/author_attestation","sign_citation":"https://pith.science/pith/GYDPBSIQ5W7FS3RIBYP677VZFS/action/citation_signature","submit_replication":"https://pith.science/pith/GYDPBSIQ5W7FS3RIBYP677VZFS/action/replication_record"}},"created_at":"2026-07-02T01:18:26.594087+00:00","updated_at":"2026-07-02T01:18:26.594087+00:00"}