{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:H6Z7WWFEHBJUJHYSEHME5BZDVO","short_pith_number":"pith:H6Z7WWFE","schema_version":"1.0","canonical_sha256":"3fb3fb58a43853449f1221d84e8723ab8379c0d65bd63ed67b586509272cb835","source":{"kind":"arxiv","id":"2501.17104","version":1},"attestation_state":"computed","paper":{"title":"COS(M+O)S: Curiosity and RL-Enhanced MCTS for Exploring Story Space via Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Tobias Materzok","submitted_at":"2025-01-28T17:44:04Z","abstract_excerpt":"We present COS(M+O)S, a System 2-inspired framework for open-ended plot development that systematically explores the vast space of possible story expansions, enabling a 3B-parameter language model to approach the plot quality of a 70B model on select short-story tasks. The method accomplishes this by combining Monte Carlo Tree Search (MCTS), guided by a step-level value model that rewards moderate surprisal (curiosity) while penalizing incoherence, and Odds Ratio Preference Optimization (ORPO) to fine-tune the policy on high-value plot expansions. This iterative reinforcement learning loop sys"},"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":"2501.17104","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-28T17:44:04Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c35215f69bfeaf81237bd49c4e0de7613c5ce87d1d1577d4e485756bd7831db6","abstract_canon_sha256":"81429f3f21b504058f910edc2b37d5a518a3898da0438c33749fa39cffa0882e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:06:34.415716Z","signature_b64":"tK4oTjMHn/LYvudA501bVzvFSvzVgEInVvnfExfFUc/6zVRg2EG4qgn5I/M7pSIcGuP7a81B3GXSVw81L6FDAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3fb3fb58a43853449f1221d84e8723ab8379c0d65bd63ed67b586509272cb835","last_reissued_at":"2026-07-05T10:06:34.415256Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:06:34.415256Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"COS(M+O)S: Curiosity and RL-Enhanced MCTS for Exploring Story Space via Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Tobias Materzok","submitted_at":"2025-01-28T17:44:04Z","abstract_excerpt":"We present COS(M+O)S, a System 2-inspired framework for open-ended plot development that systematically explores the vast space of possible story expansions, enabling a 3B-parameter language model to approach the plot quality of a 70B model on select short-story tasks. The method accomplishes this by combining Monte Carlo Tree Search (MCTS), guided by a step-level value model that rewards moderate surprisal (curiosity) while penalizing incoherence, and Odds Ratio Preference Optimization (ORPO) to fine-tune the policy on high-value plot expansions. This iterative reinforcement learning loop sys"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.17104","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/2501.17104/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":"2501.17104","created_at":"2026-07-05T10:06:34.415323+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.17104v1","created_at":"2026-07-05T10:06:34.415323+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.17104","created_at":"2026-07-05T10:06:34.415323+00:00"},{"alias_kind":"pith_short_12","alias_value":"H6Z7WWFEHBJU","created_at":"2026-07-05T10:06:34.415323+00:00"},{"alias_kind":"pith_short_16","alias_value":"H6Z7WWFEHBJUJHYS","created_at":"2026-07-05T10:06:34.415323+00:00"},{"alias_kind":"pith_short_8","alias_value":"H6Z7WWFE","created_at":"2026-07-05T10:06:34.415323+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.17323","citing_title":"A Universal Avoidance Method for Diverse Multi-branch Generation","ref_index":14,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/H6Z7WWFEHBJUJHYSEHME5BZDVO","json":"https://pith.science/pith/H6Z7WWFEHBJUJHYSEHME5BZDVO.json","graph_json":"https://pith.science/api/pith-number/H6Z7WWFEHBJUJHYSEHME5BZDVO/graph.json","events_json":"https://pith.science/api/pith-number/H6Z7WWFEHBJUJHYSEHME5BZDVO/events.json","paper":"https://pith.science/paper/H6Z7WWFE"},"agent_actions":{"view_html":"https://pith.science/pith/H6Z7WWFEHBJUJHYSEHME5BZDVO","download_json":"https://pith.science/pith/H6Z7WWFEHBJUJHYSEHME5BZDVO.json","view_paper":"https://pith.science/paper/H6Z7WWFE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.17104&json=true","fetch_graph":"https://pith.science/api/pith-number/H6Z7WWFEHBJUJHYSEHME5BZDVO/graph.json","fetch_events":"https://pith.science/api/pith-number/H6Z7WWFEHBJUJHYSEHME5BZDVO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/H6Z7WWFEHBJUJHYSEHME5BZDVO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/H6Z7WWFEHBJUJHYSEHME5BZDVO/action/storage_attestation","attest_author":"https://pith.science/pith/H6Z7WWFEHBJUJHYSEHME5BZDVO/action/author_attestation","sign_citation":"https://pith.science/pith/H6Z7WWFEHBJUJHYSEHME5BZDVO/action/citation_signature","submit_replication":"https://pith.science/pith/H6Z7WWFEHBJUJHYSEHME5BZDVO/action/replication_record"}},"created_at":"2026-07-05T10:06:34.415323+00:00","updated_at":"2026-07-05T10:06:34.415323+00:00"}