{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BNZXF5KULTWBQ6BSIJVD2JNTEY","short_pith_number":"pith:BNZXF5KU","schema_version":"1.0","canonical_sha256":"0b7372f5545cec187832426a3d25b326046d92634a3fc4b070cd5c6c194c4f19","source":{"kind":"arxiv","id":"2506.23667","version":1},"attestation_state":"computed","paper":{"title":"L0: Reinforcement Learning to Become General Agents","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jiaxing Zhang, Jingyi Xi, Junjie Zhang, Junyu Lu, Songxin Zhang, Ting Sun, Yuhua Ke, Yukun Yang, Zejian Xie, Zhuoyang Song","submitted_at":"2025-06-30T09:44:32Z","abstract_excerpt":"Training large language models (LLMs) to act as autonomous agents for multi-turn, long-horizon tasks remains significant challenges in scalability and training efficiency. To address this, we introduce L-Zero (L0), a scalable, end-to-end training pipeline for general-purpose agents. Featuring a low-cost, extensible, and sandboxed concurrent agent worker pool, L0 lowers the barrier for applying reinforcement learning in complex environments. We also introduce NB-Agent, the agent scaffold within L0, which operates in a \"code-as-action\" fashion via a Read-Eval-Print-Loop (REPL). We evaluate L0 on"},"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":"2506.23667","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-30T09:44:32Z","cross_cats_sorted":[],"title_canon_sha256":"0daaf7b6e27aa3f67e9db2cb29cde38f2f7247845aa76a6a4b875820852b1ae6","abstract_canon_sha256":"3cedaecc795614e927186c3e20dc4388efddc7bf9598b2c6dabead6443173e0e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:29:26.059614Z","signature_b64":"MTgyj1yrk3lnpNbVW6QEq32CZ+/jlTUoglaReKQWj1xOnFTv+IfLLrL8yzRe9yukvkY9WDoZtQBjgb6A+OduAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0b7372f5545cec187832426a3d25b326046d92634a3fc4b070cd5c6c194c4f19","last_reissued_at":"2026-07-05T11:29:26.059104Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:29:26.059104Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"L0: Reinforcement Learning to Become General Agents","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jiaxing Zhang, Jingyi Xi, Junjie Zhang, Junyu Lu, Songxin Zhang, Ting Sun, Yuhua Ke, Yukun Yang, Zejian Xie, Zhuoyang Song","submitted_at":"2025-06-30T09:44:32Z","abstract_excerpt":"Training large language models (LLMs) to act as autonomous agents for multi-turn, long-horizon tasks remains significant challenges in scalability and training efficiency. To address this, we introduce L-Zero (L0), a scalable, end-to-end training pipeline for general-purpose agents. Featuring a low-cost, extensible, and sandboxed concurrent agent worker pool, L0 lowers the barrier for applying reinforcement learning in complex environments. We also introduce NB-Agent, the agent scaffold within L0, which operates in a \"code-as-action\" fashion via a Read-Eval-Print-Loop (REPL). We evaluate L0 on"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.23667","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/2506.23667/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":"2506.23667","created_at":"2026-07-05T11:29:26.059163+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.23667v1","created_at":"2026-07-05T11:29:26.059163+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.23667","created_at":"2026-07-05T11:29:26.059163+00:00"},{"alias_kind":"pith_short_12","alias_value":"BNZXF5KULTWB","created_at":"2026-07-05T11:29:26.059163+00:00"},{"alias_kind":"pith_short_16","alias_value":"BNZXF5KULTWBQ6BS","created_at":"2026-07-05T11:29:26.059163+00:00"},{"alias_kind":"pith_short_8","alias_value":"BNZXF5KU","created_at":"2026-07-05T11:29:26.059163+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2603.16876","citing_title":"Multi-Modal Multi-Agent Reinforcement Learning for Radiology Report Generation","ref_index":140,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02913","citing_title":"Generate, Filter, Control, Replay: A Comprehensive Survey of Rollout Strategies for LLM Reinforcement Learning","ref_index":173,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BNZXF5KULTWBQ6BSIJVD2JNTEY","json":"https://pith.science/pith/BNZXF5KULTWBQ6BSIJVD2JNTEY.json","graph_json":"https://pith.science/api/pith-number/BNZXF5KULTWBQ6BSIJVD2JNTEY/graph.json","events_json":"https://pith.science/api/pith-number/BNZXF5KULTWBQ6BSIJVD2JNTEY/events.json","paper":"https://pith.science/paper/BNZXF5KU"},"agent_actions":{"view_html":"https://pith.science/pith/BNZXF5KULTWBQ6BSIJVD2JNTEY","download_json":"https://pith.science/pith/BNZXF5KULTWBQ6BSIJVD2JNTEY.json","view_paper":"https://pith.science/paper/BNZXF5KU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.23667&json=true","fetch_graph":"https://pith.science/api/pith-number/BNZXF5KULTWBQ6BSIJVD2JNTEY/graph.json","fetch_events":"https://pith.science/api/pith-number/BNZXF5KULTWBQ6BSIJVD2JNTEY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BNZXF5KULTWBQ6BSIJVD2JNTEY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BNZXF5KULTWBQ6BSIJVD2JNTEY/action/storage_attestation","attest_author":"https://pith.science/pith/BNZXF5KULTWBQ6BSIJVD2JNTEY/action/author_attestation","sign_citation":"https://pith.science/pith/BNZXF5KULTWBQ6BSIJVD2JNTEY/action/citation_signature","submit_replication":"https://pith.science/pith/BNZXF5KULTWBQ6BSIJVD2JNTEY/action/replication_record"}},"created_at":"2026-07-05T11:29:26.059163+00:00","updated_at":"2026-07-05T11:29:26.059163+00:00"}