{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:XHOZDMSFBC4F6BL4YQAXPYULLK","short_pith_number":"pith:XHOZDMSF","schema_version":"1.0","canonical_sha256":"b9dd91b24508b85f057cc40177e28b5aaa9ee0db35d3524b0ff37be6c88f0840","source":{"kind":"arxiv","id":"2310.08740","version":3},"attestation_state":"computed","paper":{"title":"A Zero-Shot Language Agent for Computer Control with Structured Reflection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SY","eess.SY"],"primary_cat":"cs.CL","authors_text":"Bryan Wang, Gang Li, Tao Li, Yang Li, Zhiwei Deng","submitted_at":"2023-10-12T21:53:37Z","abstract_excerpt":"Large language models (LLMs) have shown increasing capacity at planning and executing a high-level goal in a live computer environment (e.g. MiniWoB++). To perform a task, recent works often require a model to learn from trace examples of the task via either supervised learning or few/many-shot prompting. Without these trace examples, it remains a challenge how an agent can autonomously learn and improve its control on a computer, which limits the ability of an agent to perform a new task. We approach this problem with a zero-shot agent that requires no given expert traces. Our agent plans for"},"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":"2310.08740","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-10-12T21:53:37Z","cross_cats_sorted":["cs.SY","eess.SY"],"title_canon_sha256":"bbe9fda96ea73fc25a467671c6dbec6cc0218a66c407a9591bc6ea41b70f2eb0","abstract_canon_sha256":"085200c71d829984137d99234d25a781f06aeb3e30041329cef5c7d0ac353a33"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:04:01.096475Z","signature_b64":"Ofuod+l0/XKnnAFQku7jh0BeKUFoKxKZidH0W9GrVRG17idKfff1+cw6MV1T9mYjEBhni6IKpefUvhkjVFSZCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b9dd91b24508b85f057cc40177e28b5aaa9ee0db35d3524b0ff37be6c88f0840","last_reissued_at":"2026-07-05T07:04:01.095937Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:04:01.095937Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Zero-Shot Language Agent for Computer Control with Structured Reflection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SY","eess.SY"],"primary_cat":"cs.CL","authors_text":"Bryan Wang, Gang Li, Tao Li, Yang Li, Zhiwei Deng","submitted_at":"2023-10-12T21:53:37Z","abstract_excerpt":"Large language models (LLMs) have shown increasing capacity at planning and executing a high-level goal in a live computer environment (e.g. MiniWoB++). To perform a task, recent works often require a model to learn from trace examples of the task via either supervised learning or few/many-shot prompting. Without these trace examples, it remains a challenge how an agent can autonomously learn and improve its control on a computer, which limits the ability of an agent to perform a new task. We approach this problem with a zero-shot agent that requires no given expert traces. Our agent plans for"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.08740","kind":"arxiv","version":3},"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/2310.08740/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":"2310.08740","created_at":"2026-07-05T07:04:01.095996+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.08740v3","created_at":"2026-07-05T07:04:01.095996+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.08740","created_at":"2026-07-05T07:04:01.095996+00:00"},{"alias_kind":"pith_short_12","alias_value":"XHOZDMSFBC4F","created_at":"2026-07-05T07:04:01.095996+00:00"},{"alias_kind":"pith_short_16","alias_value":"XHOZDMSFBC4F6BL4","created_at":"2026-07-05T07:04:01.095996+00:00"},{"alias_kind":"pith_short_8","alias_value":"XHOZDMSF","created_at":"2026-07-05T07:04:01.095996+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.28526","citing_title":"A French OSCE Dialogue Dataset and Controllable Virtual Patient System for Clinical Training","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2412.04454","citing_title":"Aguvis: Unified Pure Vision Agents for Autonomous GUI Interaction","ref_index":89,"is_internal_anchor":false},{"citing_arxiv_id":"2510.24168","citing_title":"MGA: Memory-Driven GUI Agent for Observation-Centric Interaction","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06365","citing_title":"From Agent Loops to Deterministic Graphs: Execution Lineage for Reproducible AI-Native Work","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21375","citing_title":"VLAA-GUI: Knowing When to Stop, Recover, and Search, A Modular Framework for GUI Automation","ref_index":38,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XHOZDMSFBC4F6BL4YQAXPYULLK","json":"https://pith.science/pith/XHOZDMSFBC4F6BL4YQAXPYULLK.json","graph_json":"https://pith.science/api/pith-number/XHOZDMSFBC4F6BL4YQAXPYULLK/graph.json","events_json":"https://pith.science/api/pith-number/XHOZDMSFBC4F6BL4YQAXPYULLK/events.json","paper":"https://pith.science/paper/XHOZDMSF"},"agent_actions":{"view_html":"https://pith.science/pith/XHOZDMSFBC4F6BL4YQAXPYULLK","download_json":"https://pith.science/pith/XHOZDMSFBC4F6BL4YQAXPYULLK.json","view_paper":"https://pith.science/paper/XHOZDMSF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.08740&json=true","fetch_graph":"https://pith.science/api/pith-number/XHOZDMSFBC4F6BL4YQAXPYULLK/graph.json","fetch_events":"https://pith.science/api/pith-number/XHOZDMSFBC4F6BL4YQAXPYULLK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XHOZDMSFBC4F6BL4YQAXPYULLK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XHOZDMSFBC4F6BL4YQAXPYULLK/action/storage_attestation","attest_author":"https://pith.science/pith/XHOZDMSFBC4F6BL4YQAXPYULLK/action/author_attestation","sign_citation":"https://pith.science/pith/XHOZDMSFBC4F6BL4YQAXPYULLK/action/citation_signature","submit_replication":"https://pith.science/pith/XHOZDMSFBC4F6BL4YQAXPYULLK/action/replication_record"}},"created_at":"2026-07-05T07:04:01.095996+00:00","updated_at":"2026-07-05T07:04:01.095996+00:00"}