{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:JLQU7SVEBQXVMED4TOJ2L4AIZT","short_pith_number":"pith:JLQU7SVE","canonical_record":{"source":{"id":"2605.12913","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T02:40:28Z","cross_cats_sorted":[],"title_canon_sha256":"766daa8d1047e84f86de4b371a86636845d7f49480e0028602a03dcd011c9521","abstract_canon_sha256":"7444eb4d392079d6da7cbd69c325f4d369540ce2a50f21a9d27125570f0d9c2d"},"schema_version":"1.0"},"canonical_sha256":"4ae14fcaa40c2f56107c9b93a5f008ccfecfc096119f3a1b84becc00c187ca59","source":{"kind":"arxiv","id":"2605.12913","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.12913","created_at":"2026-05-18T03:09:10Z"},{"alias_kind":"arxiv_version","alias_value":"2605.12913v1","created_at":"2026-05-18T03:09:10Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.12913","created_at":"2026-05-18T03:09:10Z"},{"alias_kind":"pith_short_12","alias_value":"JLQU7SVEBQXV","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"JLQU7SVEBQXVMED4","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"JLQU7SVE","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:JLQU7SVEBQXVMED4TOJ2L4AIZT","target":"record","payload":{"canonical_record":{"source":{"id":"2605.12913","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T02:40:28Z","cross_cats_sorted":[],"title_canon_sha256":"766daa8d1047e84f86de4b371a86636845d7f49480e0028602a03dcd011c9521","abstract_canon_sha256":"7444eb4d392079d6da7cbd69c325f4d369540ce2a50f21a9d27125570f0d9c2d"},"schema_version":"1.0"},"canonical_sha256":"4ae14fcaa40c2f56107c9b93a5f008ccfecfc096119f3a1b84becc00c187ca59","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T03:09:10.428898Z","signature_b64":"HSBSvNMqYr3DPYMCnz387jVbh8OCiSALgffPctQMU+RGDaJmQDOZcCu4oV+V9QPnRZTNgQ5CSFp3SfNO0x7tAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4ae14fcaa40c2f56107c9b93a5f008ccfecfc096119f3a1b84becc00c187ca59","last_reissued_at":"2026-05-18T03:09:10.428168Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T03:09:10.428168Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.12913","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T03:09:10Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"VEh5yQ+Ye1+ocDyKXtBdMpFREgflOw80OeoqcbOn4AiGS5pAHH1StEHY3f6zdMxRQ8FMEWtZRialvdrFuo3PAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-23T03:05:42.541920Z"},"content_sha256":"317b164d4c172dde5666a492b1eb53b6a803e76df6032a2e5f64f9ef2f231d4f","schema_version":"1.0","event_id":"sha256:317b164d4c172dde5666a492b1eb53b6a803e76df6032a2e5f64f9ef2f231d4f"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:JLQU7SVEBQXVMED4TOJ2L4AIZT","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Revisiting DAgger in the Era of LLM-Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"DAgger with turn-level interpolation mitigates covariate shift in multi-turn LLM agents while retaining dense teacher supervision.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bo Dai, Changhao Li, Chao Zhang, Chenxiao Gao, Jiawei Huang, Niao He, Rushi Qiang","submitted_at":"2026-05-13T02:40:28Z","abstract_excerpt":"Long-horizon LM agents learn from multi-turn interaction, where a single early mistake can alter the subsequent state distribution and derail the whole trajectory. Existing recipes fall short in complementary ways: supervised fine-tuning provides dense teacher supervision but suffers from covariate shift because it is trained on off-policy teacher trajectories; while reinforcement learning with verifiable rewards avoids this off-policy mismatch by learning from on-policy rollouts but with only sparse outcome feedback. We address this dilemma by revisiting Dataset Aggregation (DAgger) for multi"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"DAgger-style training with turn-level interpolation of student and teacher policies mitigates covariate shift while retaining dense teacher supervision, producing +3.9 and +3.6 point gains on SWE-bench Verified for 4B and 8B models respectively.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That a reliable teacher policy remains available and affordable to query at every training step and that the environment state can be reset or continued after mixed student-teacher actions without introducing new distribution shifts.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"DAgger-style training with turn-level policy interpolation raises 4B and 8B LLM agents to 27.3% and 29.8% on SWE-bench Verified, beating several larger published systems.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"DAgger with turn-level interpolation mitigates covariate shift in multi-turn LLM agents while retaining dense teacher supervision.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"7a96e2bd3b01fb21b2f9c13017685477ff1ee08735f5dbd980b0fafb57e16a5b"},"source":{"id":"2605.12913","kind":"arxiv","version":1},"verdict":{"id":"a3874483-226c-44a6-9fdc-986324510ae4","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:50:59.965269Z","strongest_claim":"DAgger-style training with turn-level interpolation of student and teacher policies mitigates covariate shift while retaining dense teacher supervision, producing +3.9 and +3.6 point gains on SWE-bench Verified for 4B and 8B models respectively.","one_line_summary":"DAgger-style training with turn-level policy interpolation raises 4B and 8B LLM agents to 27.3% and 29.8% on SWE-bench Verified, beating several larger published systems.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That a reliable teacher policy remains available and affordable to query at every training step and that the environment state can be reset or continued after mixed student-teacher actions without introducing new distribution shifts.","pith_extraction_headline":"DAgger with turn-level interpolation mitigates covariate shift in multi-turn LLM agents while retaining dense teacher supervision."},"references":{"count":61,"sample":[{"doi":"","year":2023,"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","ref_index":1,"cited_arxiv_id":"2303.08774","is_internal_anchor":true},{"doi":"","year":2024,"title":"On-policy distillation of language models: Learning from self-generated mistakes","work_id":"3733ff2d-cd95-4776-9fa9-1b2328326749","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2026,"title":"Dream: Deep research evaluation with agentic metrics.arXiv preprint arXiv:2602.18940, 2026","work_id":"8010f1a4-ea3c-4f11-8c31-857e66722934","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Swe-rebench: An automated pipeline for task collection and decontaminated evaluation of software engineering agents","work_id":"c819f5fe-32b6-41d1-aeb5-2d80c6fa8474","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2026,"title":"Qwen3-Coder-Next Technical Report","work_id":"ad966e68-641d-4b33-a9da-57cf741f35a6","ref_index":5,"cited_arxiv_id":"2603.00729","is_internal_anchor":true}],"resolved_work":61,"snapshot_sha256":"0c51235a8f3e9b1cfe5acdfae95e90eaa98571c42bcbc7d6544e0ec802102218","internal_anchors":22},"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"},"verdict_id":"a3874483-226c-44a6-9fdc-986324510ae4"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T03:09:10Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"H6DJs9qkaJPdsyTrew80tkk+4DH3BSR/Xrn6ZgqoG4Zie0oLjC6OMgaHRdzEQDh0toPshaxghp/Rg5HOx204AQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-23T03:05:42.542910Z"},"content_sha256":"d9f8d3ee3c8d9d3149287bffdf2150237baf18d015ed527572b5ac07f8052336","schema_version":"1.0","event_id":"sha256:d9f8d3ee3c8d9d3149287bffdf2150237baf18d015ed527572b5ac07f8052336"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/JLQU7SVEBQXVMED4TOJ2L4AIZT/bundle.json","state_url":"https://pith.science/pith/JLQU7SVEBQXVMED4TOJ2L4AIZT/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/JLQU7SVEBQXVMED4TOJ2L4AIZT/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-23T03:05:42Z","links":{"resolver":"https://pith.science/pith/JLQU7SVEBQXVMED4TOJ2L4AIZT","bundle":"https://pith.science/pith/JLQU7SVEBQXVMED4TOJ2L4AIZT/bundle.json","state":"https://pith.science/pith/JLQU7SVEBQXVMED4TOJ2L4AIZT/state.json","well_known_bundle":"https://pith.science/.well-known/pith/JLQU7SVEBQXVMED4TOJ2L4AIZT/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:JLQU7SVEBQXVMED4TOJ2L4AIZT","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"7444eb4d392079d6da7cbd69c325f4d369540ce2a50f21a9d27125570f0d9c2d","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T02:40:28Z","title_canon_sha256":"766daa8d1047e84f86de4b371a86636845d7f49480e0028602a03dcd011c9521"},"schema_version":"1.0","source":{"id":"2605.12913","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.12913","created_at":"2026-05-18T03:09:10Z"},{"alias_kind":"arxiv_version","alias_value":"2605.12913v1","created_at":"2026-05-18T03:09:10Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.12913","created_at":"2026-05-18T03:09:10Z"},{"alias_kind":"pith_short_12","alias_value":"JLQU7SVEBQXV","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"JLQU7SVEBQXVMED4","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"JLQU7SVE","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:d9f8d3ee3c8d9d3149287bffdf2150237baf18d015ed527572b5ac07f8052336","target":"graph","created_at":"2026-05-18T03:09:10Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"DAgger-style training with turn-level interpolation of student and teacher policies mitigates covariate shift while retaining dense teacher supervision, producing +3.9 and +3.6 point gains on SWE-bench Verified for 4B and 8B models respectively."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That a reliable teacher policy remains available and affordable to query at every training step and that the environment state can be reset or continued after mixed student-teacher actions without introducing new distribution shifts."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"DAgger-style training with turn-level policy interpolation raises 4B and 8B LLM agents to 27.3% and 29.8% on SWE-bench Verified, beating several larger published systems."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"DAgger with turn-level interpolation mitigates covariate shift in multi-turn LLM agents while retaining dense teacher supervision."}],"snapshot_sha256":"7a96e2bd3b01fb21b2f9c13017685477ff1ee08735f5dbd980b0fafb57e16a5b"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Long-horizon LM agents learn from multi-turn interaction, where a single early mistake can alter the subsequent state distribution and derail the whole trajectory. Existing recipes fall short in complementary ways: supervised fine-tuning provides dense teacher supervision but suffers from covariate shift because it is trained on off-policy teacher trajectories; while reinforcement learning with verifiable rewards avoids this off-policy mismatch by learning from on-policy rollouts but with only sparse outcome feedback. We address this dilemma by revisiting Dataset Aggregation (DAgger) for multi","authors_text":"Bo Dai, Changhao Li, Chao Zhang, Chenxiao Gao, Jiawei Huang, Niao He, Rushi Qiang","cross_cats":[],"headline":"DAgger with turn-level interpolation mitigates covariate shift in multi-turn LLM agents while retaining dense teacher supervision.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T02:40:28Z","title":"Revisiting DAgger in the Era of LLM-Agents"},"references":{"count":61,"internal_anchors":22,"resolved_work":61,"sample":[{"cited_arxiv_id":"2303.08774","doi":"","is_internal_anchor":true,"ref_index":1,"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"On-policy distillation of language models: Learning from self-generated mistakes","work_id":"3733ff2d-cd95-4776-9fa9-1b2328326749","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Dream: Deep research evaluation with agentic metrics.arXiv preprint arXiv:2602.18940, 2026","work_id":"8010f1a4-ea3c-4f11-8c31-857e66722934","year":2026},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Swe-rebench: An automated pipeline for task collection and decontaminated evaluation of software engineering agents","work_id":"c819f5fe-32b6-41d1-aeb5-2d80c6fa8474","year":2025},{"cited_arxiv_id":"2603.00729","doi":"","is_internal_anchor":true,"ref_index":5,"title":"Qwen3-Coder-Next Technical Report","work_id":"ad966e68-641d-4b33-a9da-57cf741f35a6","year":2026}],"snapshot_sha256":"0c51235a8f3e9b1cfe5acdfae95e90eaa98571c42bcbc7d6544e0ec802102218"},"source":{"id":"2605.12913","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-14T19:50:59.965269Z","id":"a3874483-226c-44a6-9fdc-986324510ae4","model_set":{"reader":"grok-4.3"},"one_line_summary":"DAgger-style training with turn-level policy interpolation raises 4B and 8B LLM agents to 27.3% and 29.8% on SWE-bench Verified, beating several larger published systems.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"DAgger with turn-level interpolation mitigates covariate shift in multi-turn LLM agents while retaining dense teacher supervision.","strongest_claim":"DAgger-style training with turn-level interpolation of student and teacher policies mitigates covariate shift while retaining dense teacher supervision, producing +3.9 and +3.6 point gains on SWE-bench Verified for 4B and 8B models respectively.","weakest_assumption":"That a reliable teacher policy remains available and affordable to query at every training step and that the environment state can be reset or continued after mixed student-teacher actions without introducing new distribution shifts."}},"verdict_id":"a3874483-226c-44a6-9fdc-986324510ae4"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:317b164d4c172dde5666a492b1eb53b6a803e76df6032a2e5f64f9ef2f231d4f","target":"record","created_at":"2026-05-18T03:09:10Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"7444eb4d392079d6da7cbd69c325f4d369540ce2a50f21a9d27125570f0d9c2d","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T02:40:28Z","title_canon_sha256":"766daa8d1047e84f86de4b371a86636845d7f49480e0028602a03dcd011c9521"},"schema_version":"1.0","source":{"id":"2605.12913","kind":"arxiv","version":1}},"canonical_sha256":"4ae14fcaa40c2f56107c9b93a5f008ccfecfc096119f3a1b84becc00c187ca59","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4ae14fcaa40c2f56107c9b93a5f008ccfecfc096119f3a1b84becc00c187ca59","first_computed_at":"2026-05-18T03:09:10.428168Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T03:09:10.428168Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"HSBSvNMqYr3DPYMCnz387jVbh8OCiSALgffPctQMU+RGDaJmQDOZcCu4oV+V9QPnRZTNgQ5CSFp3SfNO0x7tAg==","signature_status":"signed_v1","signed_at":"2026-05-18T03:09:10.428898Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.12913","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:317b164d4c172dde5666a492b1eb53b6a803e76df6032a2e5f64f9ef2f231d4f","sha256:d9f8d3ee3c8d9d3149287bffdf2150237baf18d015ed527572b5ac07f8052336"],"state_sha256":"2ca66a9bd35eaf8a7057defbf511723c71b0198a1d3c9bb9a5bb822dbe4cc1ad"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"H/afAwtNISDZKpSrGB6YCrkXbXWolWJyqTSNiCJAxPU7kigXAbDWUxFNGgKet7/AczdH7p3YX5oi4PV1ziw6Aw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-23T03:05:42.547259Z","bundle_sha256":"64bfef1c9d9265d8376a409c88eb51c4fded4d18dd2f4ea3a129344ff4988abd"}}