{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:DHFJL7GBMPQOVHTGB33QBZ5S65","short_pith_number":"pith:DHFJL7GB","canonical_record":{"source":{"id":"2502.14669","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-02-20T16:05:18Z","cross_cats_sorted":[],"title_canon_sha256":"2ee1723fa380b6d3a0a0e4c95ce85f86e15b3dd06a48b97a115c0d0895ad8e2a","abstract_canon_sha256":"971b17eac94525f4287fdb302b7febadb916c8b316f4326ec77a1ef8b2d4d6d7"},"schema_version":"1.0"},"canonical_sha256":"19ca95fcc163e0ea9e660ef700e7b2f772ef49cb6b638e8384b3014997cd5d1e","source":{"kind":"arxiv","id":"2502.14669","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.14669","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"arxiv_version","alias_value":"2502.14669v3","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.14669","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"pith_short_12","alias_value":"DHFJL7GBMPQO","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"pith_short_16","alias_value":"DHFJL7GBMPQOVHTG","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"pith_short_8","alias_value":"DHFJL7GB","created_at":"2026-07-05T10:19:39Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:DHFJL7GBMPQOVHTGB33QBZ5S65","target":"record","payload":{"canonical_record":{"source":{"id":"2502.14669","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-02-20T16:05:18Z","cross_cats_sorted":[],"title_canon_sha256":"2ee1723fa380b6d3a0a0e4c95ce85f86e15b3dd06a48b97a115c0d0895ad8e2a","abstract_canon_sha256":"971b17eac94525f4287fdb302b7febadb916c8b316f4326ec77a1ef8b2d4d6d7"},"schema_version":"1.0"},"canonical_sha256":"19ca95fcc163e0ea9e660ef700e7b2f772ef49cb6b638e8384b3014997cd5d1e","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:19:39.721065Z","signature_b64":"MYuhkHun00iQXOb/ILOB3OAkppU6zYTamSxcapKG51XvKY4YBUS6F9RGvMJyrpRbA5HRccWQzMtnURT52I4XAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"19ca95fcc163e0ea9e660ef700e7b2f772ef49cb6b638e8384b3014997cd5d1e","last_reissued_at":"2026-07-05T10:19:39.720587Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:19:39.720587Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2502.14669","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:19:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"T+zQKGGCseJ2fNmMC07lzKQ9mt5Dki7rFk2Dxu/hJB6rFVfH27ROiLSbMwcNw0EKQyUoCQRntjrpxOblp5N6CA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T18:02:49.995498Z"},"content_sha256":"18d1b556133fdb68adc3e530a00c817a2d4949dd675af591adda6bcfaf1f2c99","schema_version":"1.0","event_id":"sha256:18d1b556133fdb68adc3e530a00c817a2d4949dd675af591adda6bcfaf1f2c99"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:DHFJL7GBMPQOVHTGB33QBZ5S65","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"AlphaMaze: Enhancing Large Language Models' Spatial Intelligence via GRPO","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alan Dao (Gia Tuan Dao), Dinh Bach Vu","submitted_at":"2025-02-20T16:05:18Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated impressive capabilities in language processing, yet they often struggle with tasks requiring genuine visual spatial reasoning. In this paper, we introduce a novel two-stage training framework designed to equip standard LLMs with visual reasoning abilities for maze navigation. First, we leverage Supervised Fine Tuning (SFT) on a curated dataset of tokenized maze representations to teach the model to predict step-by-step movement commands. Next, we apply Group Relative Policy Optimization (GRPO)-a technique used in DeepSeekR1-with a carefully crafte"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.14669","kind":"arxiv","version":3},"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/2502.14669/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:19:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"lmx0Asu4MJ0MyaGCJG0vaG71f4HIzuAX6MfZS7BPOsH1IUoywFVEd7Pzg51yy6DTcuBUeuBwFBe3NqrxwyCrCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T18:02:49.996151Z"},"content_sha256":"462376ea03c95caf80f3d53e0880090787a08519ed3fe586267c8ee5647143f4","schema_version":"1.0","event_id":"sha256:462376ea03c95caf80f3d53e0880090787a08519ed3fe586267c8ee5647143f4"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DHFJL7GBMPQOVHTGB33QBZ5S65/bundle.json","state_url":"https://pith.science/pith/DHFJL7GBMPQOVHTGB33QBZ5S65/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DHFJL7GBMPQOVHTGB33QBZ5S65/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-09T18:02:49Z","links":{"resolver":"https://pith.science/pith/DHFJL7GBMPQOVHTGB33QBZ5S65","bundle":"https://pith.science/pith/DHFJL7GBMPQOVHTGB33QBZ5S65/bundle.json","state":"https://pith.science/pith/DHFJL7GBMPQOVHTGB33QBZ5S65/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DHFJL7GBMPQOVHTGB33QBZ5S65/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:DHFJL7GBMPQOVHTGB33QBZ5S65","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"971b17eac94525f4287fdb302b7febadb916c8b316f4326ec77a1ef8b2d4d6d7","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-02-20T16:05:18Z","title_canon_sha256":"2ee1723fa380b6d3a0a0e4c95ce85f86e15b3dd06a48b97a115c0d0895ad8e2a"},"schema_version":"1.0","source":{"id":"2502.14669","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.14669","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"arxiv_version","alias_value":"2502.14669v3","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.14669","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"pith_short_12","alias_value":"DHFJL7GBMPQO","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"pith_short_16","alias_value":"DHFJL7GBMPQOVHTG","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"pith_short_8","alias_value":"DHFJL7GB","created_at":"2026-07-05T10:19:39Z"}],"graph_snapshots":[{"event_id":"sha256:462376ea03c95caf80f3d53e0880090787a08519ed3fe586267c8ee5647143f4","target":"graph","created_at":"2026-07-05T10:19:39Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2502.14669/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Models (LLMs) have demonstrated impressive capabilities in language processing, yet they often struggle with tasks requiring genuine visual spatial reasoning. In this paper, we introduce a novel two-stage training framework designed to equip standard LLMs with visual reasoning abilities for maze navigation. First, we leverage Supervised Fine Tuning (SFT) on a curated dataset of tokenized maze representations to teach the model to predict step-by-step movement commands. Next, we apply Group Relative Policy Optimization (GRPO)-a technique used in DeepSeekR1-with a carefully crafte","authors_text":"Alan Dao (Gia Tuan Dao), Dinh Bach Vu","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-02-20T16:05:18Z","title":"AlphaMaze: Enhancing Large Language Models' Spatial Intelligence via GRPO"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.14669","kind":"arxiv","version":3},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:18d1b556133fdb68adc3e530a00c817a2d4949dd675af591adda6bcfaf1f2c99","target":"record","created_at":"2026-07-05T10:19:39Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"971b17eac94525f4287fdb302b7febadb916c8b316f4326ec77a1ef8b2d4d6d7","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-02-20T16:05:18Z","title_canon_sha256":"2ee1723fa380b6d3a0a0e4c95ce85f86e15b3dd06a48b97a115c0d0895ad8e2a"},"schema_version":"1.0","source":{"id":"2502.14669","kind":"arxiv","version":3}},"canonical_sha256":"19ca95fcc163e0ea9e660ef700e7b2f772ef49cb6b638e8384b3014997cd5d1e","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"19ca95fcc163e0ea9e660ef700e7b2f772ef49cb6b638e8384b3014997cd5d1e","first_computed_at":"2026-07-05T10:19:39.720587Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:19:39.720587Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"MYuhkHun00iQXOb/ILOB3OAkppU6zYTamSxcapKG51XvKY4YBUS6F9RGvMJyrpRbA5HRccWQzMtnURT52I4XAg==","signature_status":"signed_v1","signed_at":"2026-07-05T10:19:39.721065Z","signed_message":"canonical_sha256_bytes"},"source_id":"2502.14669","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:18d1b556133fdb68adc3e530a00c817a2d4949dd675af591adda6bcfaf1f2c99","sha256:462376ea03c95caf80f3d53e0880090787a08519ed3fe586267c8ee5647143f4"],"state_sha256":"e67ea76dd6b077d91dab1b7d59b7322b66e085210c28ad8ac0e2c2d050de8f2d"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"L5QmNQwxXG9qoRpgD+mpL7jAyclaxfGYwYhqgL+vUMUqzCJNcUXga/znZwnfkUZ0z8BsSoEx86bx1Uh6N6zICg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T18:02:50.001441Z","bundle_sha256":"e55c4d64faaafac32c2be3484e3fcddc8604907f91a56d469dcc9c2d61e31f60"}}