{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:TCUVWHKY4N2OPJY2RXRET4WYFC","short_pith_number":"pith:TCUVWHKY","schema_version":"1.0","canonical_sha256":"98a95b1d58e374e7a71a8de249f2d828b7ab2dc7fccc969028e769743ef0b81d","source":{"kind":"arxiv","id":"2502.00684","version":2},"attestation_state":"computed","paper":{"title":"Compositional Concept-Based Neuron-Level Interpretability for Deep Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Hai Huang, Xingquan Zuo, Zeyu Jiang","submitted_at":"2025-02-02T06:05:49Z","abstract_excerpt":"Deep reinforcement learning (DRL) has successfully addressed many complex control problems. However, the neural networks representing policies or values remain opaque, undermining trust in high-stakes applications. While concept-based methods have shown promise in deciphering internal representations in computer vision, applying them to DRL is impeded by the absence of pre-defined semantic concepts in continuous state spaces. In this work, we propose a novel concept-based explanation framework designed to provide fine-grained, neuron-level insights into DRL models. Unlike previous approaches t"},"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":"2502.00684","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-02-02T06:05:49Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8e585c9e260fad9bec4661a7836365ad7102df319af5eec33d13479637c7d2e4","abstract_canon_sha256":"4a22a8e2f1fd7071513d6e829e9d8a17e7e5b252eda7b58b76f770d0544ead48"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-01T01:17:38.961520Z","signature_b64":"lx7N0sLO+GNeSo4cGhG7Th2Y3yQbwNesqlqvcXGUXCWkaoksW2pMKGFg+tis5y65pWWcBLTNVGV8S5+xnArkAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"98a95b1d58e374e7a71a8de249f2d828b7ab2dc7fccc969028e769743ef0b81d","last_reissued_at":"2026-07-01T01:17:38.960447Z","signature_status":"signed_v1","first_computed_at":"2026-07-01T01:17:38.960447Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Compositional Concept-Based Neuron-Level Interpretability for Deep Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Hai Huang, Xingquan Zuo, Zeyu Jiang","submitted_at":"2025-02-02T06:05:49Z","abstract_excerpt":"Deep reinforcement learning (DRL) has successfully addressed many complex control problems. However, the neural networks representing policies or values remain opaque, undermining trust in high-stakes applications. While concept-based methods have shown promise in deciphering internal representations in computer vision, applying them to DRL is impeded by the absence of pre-defined semantic concepts in continuous state spaces. In this work, we propose a novel concept-based explanation framework designed to provide fine-grained, neuron-level insights into DRL models. Unlike previous approaches t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.00684","kind":"arxiv","version":2},"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/2502.00684/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":"2502.00684","created_at":"2026-07-01T01:17:38.961057+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.00684v2","created_at":"2026-07-01T01:17:38.961057+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.00684","created_at":"2026-07-01T01:17:38.961057+00:00"},{"alias_kind":"pith_short_12","alias_value":"TCUVWHKY4N2O","created_at":"2026-07-01T01:17:38.961057+00:00"},{"alias_kind":"pith_short_16","alias_value":"TCUVWHKY4N2OPJY2","created_at":"2026-07-01T01:17:38.961057+00:00"},{"alias_kind":"pith_short_8","alias_value":"TCUVWHKY","created_at":"2026-07-01T01:17:38.961057+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/TCUVWHKY4N2OPJY2RXRET4WYFC","json":"https://pith.science/pith/TCUVWHKY4N2OPJY2RXRET4WYFC.json","graph_json":"https://pith.science/api/pith-number/TCUVWHKY4N2OPJY2RXRET4WYFC/graph.json","events_json":"https://pith.science/api/pith-number/TCUVWHKY4N2OPJY2RXRET4WYFC/events.json","paper":"https://pith.science/paper/TCUVWHKY"},"agent_actions":{"view_html":"https://pith.science/pith/TCUVWHKY4N2OPJY2RXRET4WYFC","download_json":"https://pith.science/pith/TCUVWHKY4N2OPJY2RXRET4WYFC.json","view_paper":"https://pith.science/paper/TCUVWHKY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.00684&json=true","fetch_graph":"https://pith.science/api/pith-number/TCUVWHKY4N2OPJY2RXRET4WYFC/graph.json","fetch_events":"https://pith.science/api/pith-number/TCUVWHKY4N2OPJY2RXRET4WYFC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TCUVWHKY4N2OPJY2RXRET4WYFC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TCUVWHKY4N2OPJY2RXRET4WYFC/action/storage_attestation","attest_author":"https://pith.science/pith/TCUVWHKY4N2OPJY2RXRET4WYFC/action/author_attestation","sign_citation":"https://pith.science/pith/TCUVWHKY4N2OPJY2RXRET4WYFC/action/citation_signature","submit_replication":"https://pith.science/pith/TCUVWHKY4N2OPJY2RXRET4WYFC/action/replication_record"}},"created_at":"2026-07-01T01:17:38.961057+00:00","updated_at":"2026-07-01T01:17:38.961057+00:00"}