{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:J7F5OYQ5MHRNOYL3PJQ762N64A","short_pith_number":"pith:J7F5OYQ5","schema_version":"1.0","canonical_sha256":"4fcbd7621d61e2d7617b7a61ff69bee007aae4e7dab92eec6ed8ab2c6bf44248","source":{"kind":"arxiv","id":"2602.11675","version":4},"attestation_state":"computed","paper":{"title":"Epistemic Regret Minimization: Label-Free Causal Critique Beyond Outcome Reward","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Epistemic Regret Minimization identifies causal flaws in LLM reasoning traces without ground-truth labels","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Edward Y. Chang, Longling Geng","submitted_at":"2026-02-12T07:48:21Z","abstract_excerpt":"Large language models can answer causal questions correctly for the wrong reasons. Current RL methods reward \\emph{what} a model concludes but ignore \\emph{why}, reinforcing correlational shortcuts -- a failure we call \\emph{Reward Entrenchment}. We introduce \\emph{Epistemic Regret Minimization} (\\erm), a framework that critiques the causal \\emph{structure} of a model's reasoning trace rather than its answer. Applying established causal principles, \\erm flags unexamined confounders, correlation--intervention conflation, and unchecked back-door paths from exposed reasoning traces. The framework"},"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":true},"canonical_record":{"source":{"id":"2602.11675","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-02-12T07:48:21Z","cross_cats_sorted":[],"title_canon_sha256":"dfc6af3c865c59bb01d646c95472519247da114c9af32f42b085052598999647","abstract_canon_sha256":"661e71f4dec067c55314260d28df484caf64fb8bcb838baa0a4d0a30ba4890d9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:04:23.152743Z","signature_b64":"PiPvBFQ/trn6i7mDjDCQAD4c91zpIuxm03YEQQYh3woXy8hi3He+10IvlHoSfqQ2zazq24BG9StgcYULujYgBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4fcbd7621d61e2d7617b7a61ff69bee007aae4e7dab92eec6ed8ab2c6bf44248","last_reissued_at":"2026-05-21T01:04:23.151766Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:04:23.151766Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Epistemic Regret Minimization: Label-Free Causal Critique Beyond Outcome Reward","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Epistemic Regret Minimization identifies causal flaws in LLM reasoning traces without ground-truth labels","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Edward Y. Chang, Longling Geng","submitted_at":"2026-02-12T07:48:21Z","abstract_excerpt":"Large language models can answer causal questions correctly for the wrong reasons. Current RL methods reward \\emph{what} a model concludes but ignore \\emph{why}, reinforcing correlational shortcuts -- a failure we call \\emph{Reward Entrenchment}. We introduce \\emph{Epistemic Regret Minimization} (\\erm), a framework that critiques the causal \\emph{structure} of a model's reasoning trace rather than its answer. Applying established causal principles, \\erm flags unexamined confounders, correlation--intervention conflation, and unchecked back-door paths from exposed reasoning traces. The framework"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"A separation theorem proves outcome-only RL cannot distinguish correct from flawed causal models in confounded environments, and preliminary experiments show epistemic reward carries signal where outcome reward does not.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That causal flaws are reliably identifiable and correctable from reasoning traces alone without ground-truth labels or external verifiers, and that the observed corrections on CausalT5K and CLadder generalize beyond the tested models and scenarios.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Epistemic Regret Minimization identifies causal reasoning flaws in LLMs from traces alone, corrects stubborn models where outcome-only methods fail, and is supported by a separation theorem proving outcome-only RL cannot distinguish correct from flawed causal models in confounded settings.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Epistemic Regret Minimization identifies causal flaws in LLM reasoning traces without ground-truth labels","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"3b31d6e4d0acc81bc9f0dd8e9cf8877ca93e9d0881f07e5a7cd81c242658ada2"},"source":{"id":"2602.11675","kind":"arxiv","version":4},"verdict":{"id":"c94b2c6a-f016-4dca-a646-1cde4ebdba1b","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T05:15:57.563360Z","strongest_claim":"A separation theorem proves outcome-only RL cannot distinguish correct from flawed causal models in confounded environments, and preliminary experiments show epistemic reward carries signal where outcome reward does not.","one_line_summary":"Epistemic Regret Minimization identifies causal reasoning flaws in LLMs from traces alone, corrects stubborn models where outcome-only methods fail, and is supported by a separation theorem proving outcome-only RL cannot distinguish correct from flawed causal models in confounded settings.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That causal flaws are reliably identifiable and correctable from reasoning traces alone without ground-truth labels or external verifiers, and that the observed corrections on CausalT5K and CLadder generalize beyond the tested models and scenarios.","pith_extraction_headline":"Epistemic Regret Minimization identifies causal flaws in LLM reasoning traces without ground-truth labels"},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2602.11675/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":3,"snapshot_sha256":"60cc0fac08e1c6b18781b3250b295ff749ea533a854ae105f5fb76b8d331e0f0"},"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":"2602.11675","created_at":"2026-05-21T01:04:23.151887+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.11675v4","created_at":"2026-05-21T01:04:23.151887+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.11675","created_at":"2026-05-21T01:04:23.151887+00:00"},{"alias_kind":"pith_short_12","alias_value":"J7F5OYQ5MHRN","created_at":"2026-05-21T01:04:23.151887+00:00"},{"alias_kind":"pith_short_16","alias_value":"J7F5OYQ5MHRNOYL3","created_at":"2026-05-21T01:04:23.151887+00:00"},{"alias_kind":"pith_short_8","alias_value":"J7F5OYQ5","created_at":"2026-05-21T01:04:23.151887+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":3,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/J7F5OYQ5MHRNOYL3PJQ762N64A","json":"https://pith.science/pith/J7F5OYQ5MHRNOYL3PJQ762N64A.json","graph_json":"https://pith.science/api/pith-number/J7F5OYQ5MHRNOYL3PJQ762N64A/graph.json","events_json":"https://pith.science/api/pith-number/J7F5OYQ5MHRNOYL3PJQ762N64A/events.json","paper":"https://pith.science/paper/J7F5OYQ5"},"agent_actions":{"view_html":"https://pith.science/pith/J7F5OYQ5MHRNOYL3PJQ762N64A","download_json":"https://pith.science/pith/J7F5OYQ5MHRNOYL3PJQ762N64A.json","view_paper":"https://pith.science/paper/J7F5OYQ5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.11675&json=true","fetch_graph":"https://pith.science/api/pith-number/J7F5OYQ5MHRNOYL3PJQ762N64A/graph.json","fetch_events":"https://pith.science/api/pith-number/J7F5OYQ5MHRNOYL3PJQ762N64A/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/J7F5OYQ5MHRNOYL3PJQ762N64A/action/timestamp_anchor","attest_storage":"https://pith.science/pith/J7F5OYQ5MHRNOYL3PJQ762N64A/action/storage_attestation","attest_author":"https://pith.science/pith/J7F5OYQ5MHRNOYL3PJQ762N64A/action/author_attestation","sign_citation":"https://pith.science/pith/J7F5OYQ5MHRNOYL3PJQ762N64A/action/citation_signature","submit_replication":"https://pith.science/pith/J7F5OYQ5MHRNOYL3PJQ762N64A/action/replication_record"}},"created_at":"2026-05-21T01:04:23.151887+00:00","updated_at":"2026-05-21T01:04:23.151887+00:00"}