{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4HF35XRJOEXACC5DIWL2F4BY5B","short_pith_number":"pith:4HF35XRJ","schema_version":"1.0","canonical_sha256":"e1cbbede29712e010ba34597a2f038e85341710b42c34d0baddabd6a41fcad7e","source":{"kind":"arxiv","id":"2606.02556","version":1},"attestation_state":"computed","paper":{"title":"HERO'S JOURNEY: Testing Complex Rule Induction with Text Games","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Anshun Asher Zheng, David I. Beaver, Junyi Jessy Li, Kanishka Misra","submitted_at":"2026-06-01T17:51:40Z","abstract_excerpt":"We introduce HERO'S JOURNEY, a benchmark for rule induction in goal-directed episodic tasks, where agents must infer hidden rules from demonstrations and act on them through multi-step execution. HERO'S JOURNEY covers eight tasks across attribute and procedural induction families, each with four structural rule forms, controllable lexical grounding, and identifiability conditions. Evaluating state-of-the-art LLMs, we find that models show evidence of rule induction, but the ability is limited and uneven across tasks. Meanwhile, process execution adds an execution bottleneck for models, whereas"},"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":"2606.02556","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-01T17:51:40Z","cross_cats_sorted":[],"title_canon_sha256":"e35534c72ca8997f6551ae9a1092fe45f97dfde726a24c5df5a0eef5af9e1a5e","abstract_canon_sha256":"9830ca1e471088d412e5fae78a4cb79e09c94d322e406dab1467d7248f457d3d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T03:05:09.501440Z","signature_b64":"jHyylyGfuu6k1GPzTV22dhcEhuV9n/HUo0kfopEGF8vFC7DOQkj78cMnJFlwoVeOksewSUr5qn75j2ZbrJCGAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e1cbbede29712e010ba34597a2f038e85341710b42c34d0baddabd6a41fcad7e","last_reissued_at":"2026-06-02T03:05:09.501086Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T03:05:09.501086Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HERO'S JOURNEY: Testing Complex Rule Induction with Text Games","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Anshun Asher Zheng, David I. Beaver, Junyi Jessy Li, Kanishka Misra","submitted_at":"2026-06-01T17:51:40Z","abstract_excerpt":"We introduce HERO'S JOURNEY, a benchmark for rule induction in goal-directed episodic tasks, where agents must infer hidden rules from demonstrations and act on them through multi-step execution. HERO'S JOURNEY covers eight tasks across attribute and procedural induction families, each with four structural rule forms, controllable lexical grounding, and identifiability conditions. Evaluating state-of-the-art LLMs, we find that models show evidence of rule induction, but the ability is limited and uneven across tasks. Meanwhile, process execution adds an execution bottleneck for models, whereas"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.02556","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/2606.02556/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":"2606.02556","created_at":"2026-06-02T03:05:09.501146+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.02556v1","created_at":"2026-06-02T03:05:09.501146+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.02556","created_at":"2026-06-02T03:05:09.501146+00:00"},{"alias_kind":"pith_short_12","alias_value":"4HF35XRJOEXA","created_at":"2026-06-02T03:05:09.501146+00:00"},{"alias_kind":"pith_short_16","alias_value":"4HF35XRJOEXACC5D","created_at":"2026-06-02T03:05:09.501146+00:00"},{"alias_kind":"pith_short_8","alias_value":"4HF35XRJ","created_at":"2026-06-02T03:05:09.501146+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/4HF35XRJOEXACC5DIWL2F4BY5B","json":"https://pith.science/pith/4HF35XRJOEXACC5DIWL2F4BY5B.json","graph_json":"https://pith.science/api/pith-number/4HF35XRJOEXACC5DIWL2F4BY5B/graph.json","events_json":"https://pith.science/api/pith-number/4HF35XRJOEXACC5DIWL2F4BY5B/events.json","paper":"https://pith.science/paper/4HF35XRJ"},"agent_actions":{"view_html":"https://pith.science/pith/4HF35XRJOEXACC5DIWL2F4BY5B","download_json":"https://pith.science/pith/4HF35XRJOEXACC5DIWL2F4BY5B.json","view_paper":"https://pith.science/paper/4HF35XRJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.02556&json=true","fetch_graph":"https://pith.science/api/pith-number/4HF35XRJOEXACC5DIWL2F4BY5B/graph.json","fetch_events":"https://pith.science/api/pith-number/4HF35XRJOEXACC5DIWL2F4BY5B/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4HF35XRJOEXACC5DIWL2F4BY5B/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4HF35XRJOEXACC5DIWL2F4BY5B/action/storage_attestation","attest_author":"https://pith.science/pith/4HF35XRJOEXACC5DIWL2F4BY5B/action/author_attestation","sign_citation":"https://pith.science/pith/4HF35XRJOEXACC5DIWL2F4BY5B/action/citation_signature","submit_replication":"https://pith.science/pith/4HF35XRJOEXACC5DIWL2F4BY5B/action/replication_record"}},"created_at":"2026-06-02T03:05:09.501146+00:00","updated_at":"2026-06-02T03:05:09.501146+00:00"}