{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:6XZBWBFYWI7CHOL4IFXBHHLCSF","short_pith_number":"pith:6XZBWBFY","schema_version":"1.0","canonical_sha256":"f5f21b04b8b23e23b97c416e139d62915eace739bfaad50cd5f19d2dba4d4d91","source":{"kind":"arxiv","id":"2606.01060","version":1},"attestation_state":"computed","paper":{"title":"MENTIS: What Belief Changes Under Alignment? Measuring Multi-Scale Latent Torsion in Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Aman Chadha, Amitava Das, Amit Dhanda, Mayur Parvatikar, Partha Pratim Saha, Samarth Raina, Vinija Jain","submitted_at":"2026-05-31T07:05:51Z","abstract_excerpt":"Preference alignment has substantially improved the observable behavior of large language models, yet it remains unclear what alignment changes internally. Aligned systems still fail under jailbreaks, prompt injection, and retrieval-time corruption, suggesting behavior-level evaluation alone is incomplete. Post-training should leave measurable traces in internal computation. We ask: when an instruction-tuned (IT) model becomes a preference-aligned (PA) model, what geometric structure changes, where do those changes concentrate, and how selectively do they vary across concepts, prompts, and mod"},"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":"2606.01060","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-31T07:05:51Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"01d35462e74fecac2c896d77c98d864b0db6573a39dd503310df481bb41b60ef","abstract_canon_sha256":"50b9d0798d29585914989a68a5e18672ee53f6e75a8ff30868708ecde6b029d7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T01:04:20.103186Z","signature_b64":"DKOSF0NzIJQR/bo4VL7p+2dmxhkFnwUzD92zH6781rff8zH6cs8zZ0NRptiG/oRMrs2QR6Z2aPhGARBjcA7YCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f5f21b04b8b23e23b97c416e139d62915eace739bfaad50cd5f19d2dba4d4d91","last_reissued_at":"2026-06-02T01:04:20.102728Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T01:04:20.102728Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MENTIS: What Belief Changes Under Alignment? Measuring Multi-Scale Latent Torsion in Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Aman Chadha, Amitava Das, Amit Dhanda, Mayur Parvatikar, Partha Pratim Saha, Samarth Raina, Vinija Jain","submitted_at":"2026-05-31T07:05:51Z","abstract_excerpt":"Preference alignment has substantially improved the observable behavior of large language models, yet it remains unclear what alignment changes internally. Aligned systems still fail under jailbreaks, prompt injection, and retrieval-time corruption, suggesting behavior-level evaluation alone is incomplete. Post-training should leave measurable traces in internal computation. We ask: when an instruction-tuned (IT) model becomes a preference-aligned (PA) model, what geometric structure changes, where do those changes concentrate, and how selectively do they vary across concepts, prompts, and mod"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.01060","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/2606.01060/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":"2606.01060","created_at":"2026-06-02T01:04:20.102792+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.01060v1","created_at":"2026-06-02T01:04:20.102792+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.01060","created_at":"2026-06-02T01:04:20.102792+00:00"},{"alias_kind":"pith_short_12","alias_value":"6XZBWBFYWI7C","created_at":"2026-06-02T01:04:20.102792+00:00"},{"alias_kind":"pith_short_16","alias_value":"6XZBWBFYWI7CHOL4","created_at":"2026-06-02T01:04:20.102792+00:00"},{"alias_kind":"pith_short_8","alias_value":"6XZBWBFY","created_at":"2026-06-02T01:04:20.102792+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/6XZBWBFYWI7CHOL4IFXBHHLCSF","json":"https://pith.science/pith/6XZBWBFYWI7CHOL4IFXBHHLCSF.json","graph_json":"https://pith.science/api/pith-number/6XZBWBFYWI7CHOL4IFXBHHLCSF/graph.json","events_json":"https://pith.science/api/pith-number/6XZBWBFYWI7CHOL4IFXBHHLCSF/events.json","paper":"https://pith.science/paper/6XZBWBFY"},"agent_actions":{"view_html":"https://pith.science/pith/6XZBWBFYWI7CHOL4IFXBHHLCSF","download_json":"https://pith.science/pith/6XZBWBFYWI7CHOL4IFXBHHLCSF.json","view_paper":"https://pith.science/paper/6XZBWBFY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.01060&json=true","fetch_graph":"https://pith.science/api/pith-number/6XZBWBFYWI7CHOL4IFXBHHLCSF/graph.json","fetch_events":"https://pith.science/api/pith-number/6XZBWBFYWI7CHOL4IFXBHHLCSF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6XZBWBFYWI7CHOL4IFXBHHLCSF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6XZBWBFYWI7CHOL4IFXBHHLCSF/action/storage_attestation","attest_author":"https://pith.science/pith/6XZBWBFYWI7CHOL4IFXBHHLCSF/action/author_attestation","sign_citation":"https://pith.science/pith/6XZBWBFYWI7CHOL4IFXBHHLCSF/action/citation_signature","submit_replication":"https://pith.science/pith/6XZBWBFYWI7CHOL4IFXBHHLCSF/action/replication_record"}},"created_at":"2026-06-02T01:04:20.102792+00:00","updated_at":"2026-06-02T01:04:20.102792+00:00"}