{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:HPWKLIRFEMYOACML5FHLMSASTD","short_pith_number":"pith:HPWKLIRF","schema_version":"1.0","canonical_sha256":"3beca5a2252330e0098be94eb6481298e07caad7945ab7df9824a8fa0ffe699c","source":{"kind":"arxiv","id":"2402.19446","version":1},"attestation_state":"computed","paper":{"title":"ArCHer: Training Language Model Agents via Hierarchical Multi-Turn RL","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Andrea Zanette, Aviral Kumar, Jiayi Pan, Sergey Levine, Yifei Zhou","submitted_at":"2024-02-29T18:45:56Z","abstract_excerpt":"A broad use case of large language models (LLMs) is in goal-directed decision-making tasks (or \"agent\" tasks), where an LLM needs to not just generate completions for a given prompt, but rather make intelligent decisions over a multi-turn interaction to accomplish a task (e.g., when interacting with the web, using tools, or providing customer support). Reinforcement learning (RL) provides a general paradigm to address such agent tasks, but current RL methods for LLMs largely focus on optimizing single-turn rewards. By construction, most single-turn RL methods cannot endow LLMs with the ability"},"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":"2402.19446","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-29T18:45:56Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"d1d7f17cab12ea5d98f991ec66f1b6b8e3672a4b30b44d508fec5be83ab7bb31","abstract_canon_sha256":"e2f6cb5205d544b284fe428648cb1394f4594a14b12c125721c0cfaeda0c8e61"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:50:45.373678Z","signature_b64":"PmMU4T5yUTtQ/1EAIQIhwwhXPN8qMpluiQXSgQS9UL0x6nW6Iooj5EU8PbfX/OKI32IxaH6NbTCrTnql+KxPDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3beca5a2252330e0098be94eb6481298e07caad7945ab7df9824a8fa0ffe699c","last_reissued_at":"2026-07-05T07:50:45.373181Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:50:45.373181Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ArCHer: Training Language Model Agents via Hierarchical Multi-Turn RL","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Andrea Zanette, Aviral Kumar, Jiayi Pan, Sergey Levine, Yifei Zhou","submitted_at":"2024-02-29T18:45:56Z","abstract_excerpt":"A broad use case of large language models (LLMs) is in goal-directed decision-making tasks (or \"agent\" tasks), where an LLM needs to not just generate completions for a given prompt, but rather make intelligent decisions over a multi-turn interaction to accomplish a task (e.g., when interacting with the web, using tools, or providing customer support). Reinforcement learning (RL) provides a general paradigm to address such agent tasks, but current RL methods for LLMs largely focus on optimizing single-turn rewards. By construction, most single-turn RL methods cannot endow LLMs with the ability"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.19446","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/2402.19446/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":"2402.19446","created_at":"2026-07-05T07:50:45.373241+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.19446v1","created_at":"2026-07-05T07:50:45.373241+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.19446","created_at":"2026-07-05T07:50:45.373241+00:00"},{"alias_kind":"pith_short_12","alias_value":"HPWKLIRFEMYO","created_at":"2026-07-05T07:50:45.373241+00:00"},{"alias_kind":"pith_short_16","alias_value":"HPWKLIRFEMYOACML","created_at":"2026-07-05T07:50:45.373241+00:00"},{"alias_kind":"pith_short_8","alias_value":"HPWKLIRF","created_at":"2026-07-05T07:50:45.373241+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":15,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26918","citing_title":"Diagnosing Task Insensitivity in Language Agents","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2606.21943","citing_title":"Modularized Reinforcement Learning on LLMs: From MDP Creation to Exploration and Learning","ref_index":279,"is_internal_anchor":false},{"citing_arxiv_id":"2606.05885","citing_title":"When Denser Credit Is Not Enough: Evidence-Calibrated Policy Optimization for Long-Horizon LLM Agent Training","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02355","citing_title":"SIRI: Self-Internalizing Reinforcement Learning with Intrinsic Skills for LLM Agent Training","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2606.01249","citing_title":"Trust Region On-Policy Distillation","ref_index":146,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22240","citing_title":"Unlocking Proactivity in Task-Oriented Dialogue","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2605.24517","citing_title":"ECHO: Terminal Agents Learn World Models for Free","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22240","citing_title":"Unlocking Proactivity in Task-Oriented Dialogue","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2510.13727","citing_title":"From Refusal to Recovery: A Control-Theoretic Approach to Generative AI Guardrails","ref_index":61,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20061","citing_title":"Rewarding Beliefs, Not Actions: Consistency-Guided Credit Assignment for Long-Horizon Agents","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2409.12917","citing_title":"Training Language Models to Self-Correct via Reinforcement Learning","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2501.09686","citing_title":"Towards Large Reasoning Models: A Survey of Reinforced Reasoning with Large Language Models","ref_index":200,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06642","citing_title":"StraTA: Incentivizing Agentic Reinforcement Learning with Strategic Trajectory Abstraction","ref_index":53,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05413","citing_title":"From History to State: Constant-Context Skill Learning for LLM Agents","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07725","citing_title":"SOD: Step-wise On-policy Distillation for Small Language Model Agents","ref_index":50,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HPWKLIRFEMYOACML5FHLMSASTD","json":"https://pith.science/pith/HPWKLIRFEMYOACML5FHLMSASTD.json","graph_json":"https://pith.science/api/pith-number/HPWKLIRFEMYOACML5FHLMSASTD/graph.json","events_json":"https://pith.science/api/pith-number/HPWKLIRFEMYOACML5FHLMSASTD/events.json","paper":"https://pith.science/paper/HPWKLIRF"},"agent_actions":{"view_html":"https://pith.science/pith/HPWKLIRFEMYOACML5FHLMSASTD","download_json":"https://pith.science/pith/HPWKLIRFEMYOACML5FHLMSASTD.json","view_paper":"https://pith.science/paper/HPWKLIRF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.19446&json=true","fetch_graph":"https://pith.science/api/pith-number/HPWKLIRFEMYOACML5FHLMSASTD/graph.json","fetch_events":"https://pith.science/api/pith-number/HPWKLIRFEMYOACML5FHLMSASTD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HPWKLIRFEMYOACML5FHLMSASTD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HPWKLIRFEMYOACML5FHLMSASTD/action/storage_attestation","attest_author":"https://pith.science/pith/HPWKLIRFEMYOACML5FHLMSASTD/action/author_attestation","sign_citation":"https://pith.science/pith/HPWKLIRFEMYOACML5FHLMSASTD/action/citation_signature","submit_replication":"https://pith.science/pith/HPWKLIRFEMYOACML5FHLMSASTD/action/replication_record"}},"created_at":"2026-07-05T07:50:45.373241+00:00","updated_at":"2026-07-05T07:50:45.373241+00:00"}