{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2017:YDUPI5DHDOM6X4USPMMWMVF75A","short_pith_number":"pith:YDUPI5DH","schema_version":"1.0","canonical_sha256":"c0e8f474671b99ebf2927b196654bfe81305fef03e6528d48db29c36c397b066","source":{"kind":"arxiv","id":"1711.09883","version":2},"attestation_state":"computed","paper":{"title":"AI Safety Gridworlds","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Andrew Lefrancq, Jan Leike, Laurent Orseau, Miljan Martic, Pedro A. Ortega, Shane Legg, Tom Everitt, Victoria Krakovna","submitted_at":"2017-11-27T18:57:13Z","abstract_excerpt":"We present a suite of reinforcement learning environments illustrating various safety properties of intelligent agents. These problems include safe interruptibility, avoiding side effects, absent supervisor, reward gaming, safe exploration, as well as robustness to self-modification, distributional shift, and adversaries. To measure compliance with the intended safe behavior, we equip each environment with a performance function that is hidden from the agent. This allows us to categorize AI safety problems into robustness and specification problems, depending on whether the performance functio"},"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":"1711.09883","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-11-27T18:57:13Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8c94b24b36ae763934080e879bc2af191d4dc48ae6819ab95ab5221252823db0","abstract_canon_sha256":"ff2cfaebbfd0cfe89c404a69141d3827c3e824ee381c0c213825c394ddf0268e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:29:23.903017Z","signature_b64":"vi01KZ33KI3aiExF0WG6eGxaDHT3GldhHx8+piizICCsz8O6ZbBw/mCqxGOSOnOh6qi/UTHe+7bg4p+r1t/gDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c0e8f474671b99ebf2927b196654bfe81305fef03e6528d48db29c36c397b066","last_reissued_at":"2026-05-18T00:29:23.902297Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:29:23.902297Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AI Safety Gridworlds","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Andrew Lefrancq, Jan Leike, Laurent Orseau, Miljan Martic, Pedro A. Ortega, Shane Legg, Tom Everitt, Victoria Krakovna","submitted_at":"2017-11-27T18:57:13Z","abstract_excerpt":"We present a suite of reinforcement learning environments illustrating various safety properties of intelligent agents. These problems include safe interruptibility, avoiding side effects, absent supervisor, reward gaming, safe exploration, as well as robustness to self-modification, distributional shift, and adversaries. To measure compliance with the intended safe behavior, we equip each environment with a performance function that is hidden from the agent. This allows us to categorize AI safety problems into robustness and specification problems, depending on whether the performance functio"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1711.09883","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":""},"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":"1711.09883","created_at":"2026-05-18T00:29:23.902415+00:00"},{"alias_kind":"arxiv_version","alias_value":"1711.09883v2","created_at":"2026-05-18T00:29:23.902415+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1711.09883","created_at":"2026-05-18T00:29:23.902415+00:00"},{"alias_kind":"pith_short_12","alias_value":"YDUPI5DHDOM6","created_at":"2026-05-18T12:31:56.362134+00:00"},{"alias_kind":"pith_short_16","alias_value":"YDUPI5DHDOM6X4US","created_at":"2026-05-18T12:31:56.362134+00:00"},{"alias_kind":"pith_short_8","alias_value":"YDUPI5DH","created_at":"2026-05-18T12:31:56.362134+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":13,"internal_anchor_count":7,"sample":[{"citing_arxiv_id":"1906.09136","citing_title":"Categorizing Wireheading in Partially Embedded Agents","ref_index":14,"is_internal_anchor":true},{"citing_arxiv_id":"1907.01285","citing_title":"Learning the Arrow of Time","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"1907.01475","citing_title":"Generalizing from a few environments in safety-critical reinforcement learning","ref_index":20,"is_internal_anchor":true},{"citing_arxiv_id":"1907.04534","citing_title":"The Role of Cooperation in Responsible AI Development","ref_index":22,"is_internal_anchor":true},{"citing_arxiv_id":"2503.03480","citing_title":"SafeVLA: Towards Safety Alignment of Vision-Language-Action Model via Constrained Learning","ref_index":61,"is_internal_anchor":true},{"citing_arxiv_id":"2506.17792","citing_title":"Accelerating Policy Synthesis in Large-Scale MDPs via Hierarchical Adaptive Refinement","ref_index":33,"is_internal_anchor":true},{"citing_arxiv_id":"2508.09128","citing_title":"A Review On Safe Reinforcement Learning Using Lyapunov and Barrier Functions","ref_index":128,"is_internal_anchor":true},{"citing_arxiv_id":"2605.03512","citing_title":"Brainrot: Deskilling and Addiction are Overlooked AI Risks","ref_index":60,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23210","citing_title":"Discovering Agentic Safety Specifications from 1-Bit Danger Signals","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05921","citing_title":"Intentmaking and Sensemaking: Human Interaction with AI-Guided Mathematical Discovery","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2604.12500","citing_title":"Safety Training Modulates Harmful Misalignment Under On-Policy RL, But Direction Depends on Environment Design","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07728","citing_title":"SARC: A Governance-by-Architecture Framework for Agentic AI Systems","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2604.13602","citing_title":"Reward Hacking in the Era of Large Models: Mechanisms, Emergent Misalignment, Challenges","ref_index":82,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YDUPI5DHDOM6X4USPMMWMVF75A","json":"https://pith.science/pith/YDUPI5DHDOM6X4USPMMWMVF75A.json","graph_json":"https://pith.science/api/pith-number/YDUPI5DHDOM6X4USPMMWMVF75A/graph.json","events_json":"https://pith.science/api/pith-number/YDUPI5DHDOM6X4USPMMWMVF75A/events.json","paper":"https://pith.science/paper/YDUPI5DH"},"agent_actions":{"view_html":"https://pith.science/pith/YDUPI5DHDOM6X4USPMMWMVF75A","download_json":"https://pith.science/pith/YDUPI5DHDOM6X4USPMMWMVF75A.json","view_paper":"https://pith.science/paper/YDUPI5DH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1711.09883&json=true","fetch_graph":"https://pith.science/api/pith-number/YDUPI5DHDOM6X4USPMMWMVF75A/graph.json","fetch_events":"https://pith.science/api/pith-number/YDUPI5DHDOM6X4USPMMWMVF75A/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YDUPI5DHDOM6X4USPMMWMVF75A/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YDUPI5DHDOM6X4USPMMWMVF75A/action/storage_attestation","attest_author":"https://pith.science/pith/YDUPI5DHDOM6X4USPMMWMVF75A/action/author_attestation","sign_citation":"https://pith.science/pith/YDUPI5DHDOM6X4USPMMWMVF75A/action/citation_signature","submit_replication":"https://pith.science/pith/YDUPI5DHDOM6X4USPMMWMVF75A/action/replication_record"}},"created_at":"2026-05-18T00:29:23.902415+00:00","updated_at":"2026-05-18T00:29:23.902415+00:00"}