{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:VO67JS32XCSZ2IM53PNMCLCKQM","short_pith_number":"pith:VO67JS32","schema_version":"1.0","canonical_sha256":"abbdf4cb7ab8a59d219ddbdac12c4a83093520d3bfcdbc38fb9a5bdeefae1f9e","source":{"kind":"arxiv","id":"2507.16252","version":1},"attestation_state":"computed","paper":{"title":"Efficient RL for optimizing conversation level outcomes with an LLM-based tutor","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Amy Zhang, Dean Foster, Emma Brunskill, Hyunji Nam, Lyle Ungar, Omer Gottesman","submitted_at":"2025-07-22T05:56:46Z","abstract_excerpt":"Large language models (LLMs) built on existing reinforcement learning with human feedback (RLHF) frameworks typically optimize responses based on immediate turn-level human preferences. However, this approach falls short in multi-turn dialogue settings, such as online math tutoring. We propose a method to enhance LLM-based tutors by representing the dialogue history with a lower-dimensional latent state representation of a student and optimizing a long-term policy to determine high-level actions based on the latent state. The goal is to better align the tutor's behavior with the long-term obje"},"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":"2507.16252","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CL","submitted_at":"2025-07-22T05:56:46Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"24258c4c6208f2b94e71e7ab86201a7a7cc718586c073ac4767ca79076fd42ab","abstract_canon_sha256":"989aa2f953757a344dfba66d98ab62090366a13b1c3d130ee193d70a30f2775e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:41:06.458257Z","signature_b64":"GD8VwKqizqU3ezCx9xDahoJ/TauJSvaAiUw/3O2vmaPYEEfZXBSTWD4uw/o5PFSfmawbaXOw/Q9Gbyr3ByCCBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"abbdf4cb7ab8a59d219ddbdac12c4a83093520d3bfcdbc38fb9a5bdeefae1f9e","last_reissued_at":"2026-07-05T11:41:06.457827Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:41:06.457827Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Efficient RL for optimizing conversation level outcomes with an LLM-based tutor","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Amy Zhang, Dean Foster, Emma Brunskill, Hyunji Nam, Lyle Ungar, Omer Gottesman","submitted_at":"2025-07-22T05:56:46Z","abstract_excerpt":"Large language models (LLMs) built on existing reinforcement learning with human feedback (RLHF) frameworks typically optimize responses based on immediate turn-level human preferences. However, this approach falls short in multi-turn dialogue settings, such as online math tutoring. We propose a method to enhance LLM-based tutors by representing the dialogue history with a lower-dimensional latent state representation of a student and optimizing a long-term policy to determine high-level actions based on the latent state. The goal is to better align the tutor's behavior with the long-term obje"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.16252","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/2507.16252/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":"2507.16252","created_at":"2026-07-05T11:41:06.457880+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.16252v1","created_at":"2026-07-05T11:41:06.457880+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.16252","created_at":"2026-07-05T11:41:06.457880+00:00"},{"alias_kind":"pith_short_12","alias_value":"VO67JS32XCSZ","created_at":"2026-07-05T11:41:06.457880+00:00"},{"alias_kind":"pith_short_16","alias_value":"VO67JS32XCSZ2IM5","created_at":"2026-07-05T11:41:06.457880+00:00"},{"alias_kind":"pith_short_8","alias_value":"VO67JS32","created_at":"2026-07-05T11:41:06.457880+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.11744","citing_title":"Hey Chat, Can You Teach Me? Structuring Socratic Dialogue for Human Learning in the Wild","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2512.20773","citing_title":"DIAL: Direct Iterative Adversarial Learning for Realistic Multi-Turn Dialogue Simulation","ref_index":1,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VO67JS32XCSZ2IM53PNMCLCKQM","json":"https://pith.science/pith/VO67JS32XCSZ2IM53PNMCLCKQM.json","graph_json":"https://pith.science/api/pith-number/VO67JS32XCSZ2IM53PNMCLCKQM/graph.json","events_json":"https://pith.science/api/pith-number/VO67JS32XCSZ2IM53PNMCLCKQM/events.json","paper":"https://pith.science/paper/VO67JS32"},"agent_actions":{"view_html":"https://pith.science/pith/VO67JS32XCSZ2IM53PNMCLCKQM","download_json":"https://pith.science/pith/VO67JS32XCSZ2IM53PNMCLCKQM.json","view_paper":"https://pith.science/paper/VO67JS32","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.16252&json=true","fetch_graph":"https://pith.science/api/pith-number/VO67JS32XCSZ2IM53PNMCLCKQM/graph.json","fetch_events":"https://pith.science/api/pith-number/VO67JS32XCSZ2IM53PNMCLCKQM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VO67JS32XCSZ2IM53PNMCLCKQM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VO67JS32XCSZ2IM53PNMCLCKQM/action/storage_attestation","attest_author":"https://pith.science/pith/VO67JS32XCSZ2IM53PNMCLCKQM/action/author_attestation","sign_citation":"https://pith.science/pith/VO67JS32XCSZ2IM53PNMCLCKQM/action/citation_signature","submit_replication":"https://pith.science/pith/VO67JS32XCSZ2IM53PNMCLCKQM/action/replication_record"}},"created_at":"2026-07-05T11:41:06.457880+00:00","updated_at":"2026-07-05T11:41:06.457880+00:00"}