{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:EDSDDNC5XDA2MRS7FO4HOLCLDN","short_pith_number":"pith:EDSDDNC5","schema_version":"1.0","canonical_sha256":"20e431b45db8c1a6465f2bb8772c4b1b7b08e786785e81eff46c698cea19d14f","source":{"kind":"arxiv","id":"2502.06776","version":2},"attestation_state":"computed","paper":{"title":"InSTA: Towards Internet-Scale Training For Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Brandon Trabucco, Gunnar Sigurdsson, Robinson Piramuthu, Ruslan Salakhutdinov","submitted_at":"2025-02-10T18:54:05Z","abstract_excerpt":"The predominant approach for training web navigation agents is to gather human demonstrations for a set of popular websites and hand-written tasks, but it is becoming clear that human data is an inefficient resource. We develop a pipeline to facilitate internet-scale training for agents without laborious human annotations. In the first stage, an LLM annotates 150k sites with agentic tasks. In the next stage, LLM agents complete tasks and produce trajectories. In the final stage, an LLM filters trajectories by judging their success. Language models are powerful data curation tools, identifying "},"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":"2502.06776","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-02-10T18:54:05Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"924396e19f7db63a085677ce7e9a1ced988b2c6bf4356b85ed5a39f0b0014556","abstract_canon_sha256":"c9d1733c5e0e43604cb8b318d416521eaf147a6359592453adff64aa61c2f2db"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:07:20.249766Z","signature_b64":"DjQKrVHqoXr3ti08GU5PvxnGNN0MzaYdwOUURs1S0AA3+npvi6Bq5CNfXSXnR2mfXn9NreO2mqA3zztH3FOdCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"20e431b45db8c1a6465f2bb8772c4b1b7b08e786785e81eff46c698cea19d14f","last_reissued_at":"2026-07-05T11:07:20.249293Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:07:20.249293Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"InSTA: Towards Internet-Scale Training For Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Brandon Trabucco, Gunnar Sigurdsson, Robinson Piramuthu, Ruslan Salakhutdinov","submitted_at":"2025-02-10T18:54:05Z","abstract_excerpt":"The predominant approach for training web navigation agents is to gather human demonstrations for a set of popular websites and hand-written tasks, but it is becoming clear that human data is an inefficient resource. We develop a pipeline to facilitate internet-scale training for agents without laborious human annotations. In the first stage, an LLM annotates 150k sites with agentic tasks. In the next stage, LLM agents complete tasks and produce trajectories. In the final stage, an LLM filters trajectories by judging their success. Language models are powerful data curation tools, identifying "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.06776","kind":"arxiv","version":2},"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/2502.06776/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":"2502.06776","created_at":"2026-07-05T11:07:20.249355+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.06776v2","created_at":"2026-07-05T11:07:20.249355+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.06776","created_at":"2026-07-05T11:07:20.249355+00:00"},{"alias_kind":"pith_short_12","alias_value":"EDSDDNC5XDA2","created_at":"2026-07-05T11:07:20.249355+00:00"},{"alias_kind":"pith_short_16","alias_value":"EDSDDNC5XDA2MRS7","created_at":"2026-07-05T11:07:20.249355+00:00"},{"alias_kind":"pith_short_8","alias_value":"EDSDDNC5","created_at":"2026-07-05T11:07:20.249355+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":8,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.27330","citing_title":"Empowering GUI Agents via Autonomous Experience Exploration and Hindsight Experience Utilization for Task Planning","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2606.17321","citing_title":"ProCUA-SFT Technical Report","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2606.12191","citing_title":"Agentic Environment Engineering for Large Language Models: A Survey of Environment Modeling, Synthesis, Evaluation, and Application","ref_index":267,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02031","citing_title":"OpenWebRL: Demystifying Online Multi-turn Reinforcement Learning for Visual Web Agents","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2601.22149","citing_title":"DynaWeb: Model-Based Reinforcement Learning of Web Agents","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2603.05044","citing_title":"WebFactory: Automated Compression of Foundational Language Intelligence into Grounded Web Agents","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2603.20340","citing_title":"ContractSkill: Repairable Contract-Based Skills for Multimodal Web Agents","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06761","citing_title":"Weblica: Scalable and Reproducible Training Environments for Visual Web Agents","ref_index":33,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EDSDDNC5XDA2MRS7FO4HOLCLDN","json":"https://pith.science/pith/EDSDDNC5XDA2MRS7FO4HOLCLDN.json","graph_json":"https://pith.science/api/pith-number/EDSDDNC5XDA2MRS7FO4HOLCLDN/graph.json","events_json":"https://pith.science/api/pith-number/EDSDDNC5XDA2MRS7FO4HOLCLDN/events.json","paper":"https://pith.science/paper/EDSDDNC5"},"agent_actions":{"view_html":"https://pith.science/pith/EDSDDNC5XDA2MRS7FO4HOLCLDN","download_json":"https://pith.science/pith/EDSDDNC5XDA2MRS7FO4HOLCLDN.json","view_paper":"https://pith.science/paper/EDSDDNC5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.06776&json=true","fetch_graph":"https://pith.science/api/pith-number/EDSDDNC5XDA2MRS7FO4HOLCLDN/graph.json","fetch_events":"https://pith.science/api/pith-number/EDSDDNC5XDA2MRS7FO4HOLCLDN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EDSDDNC5XDA2MRS7FO4HOLCLDN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EDSDDNC5XDA2MRS7FO4HOLCLDN/action/storage_attestation","attest_author":"https://pith.science/pith/EDSDDNC5XDA2MRS7FO4HOLCLDN/action/author_attestation","sign_citation":"https://pith.science/pith/EDSDDNC5XDA2MRS7FO4HOLCLDN/action/citation_signature","submit_replication":"https://pith.science/pith/EDSDDNC5XDA2MRS7FO4HOLCLDN/action/replication_record"}},"created_at":"2026-07-05T11:07:20.249355+00:00","updated_at":"2026-07-05T11:07:20.249355+00:00"}