{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:3Z6ZFPYJIGTZL23JGJIZYXC6DR","short_pith_number":"pith:3Z6ZFPYJ","canonical_record":{"source":{"id":"2306.13649","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-06-23T17:56:26Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"a1e839e0ebb01d6ed89cbf56a10842de6baaa7a05209b87ef5f411622e0fe35d","abstract_canon_sha256":"389ee6658f57e9b1a62394a4d26a57ea927da59024257f05bf61fe567e9be085"},"schema_version":"1.0"},"canonical_sha256":"de7d92bf0941a795eb6932519c5c5e1c5d1e0d359bdb17e8871a4d3f559c3fb6","source":{"kind":"arxiv","id":"2306.13649","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2306.13649","created_at":"2026-07-05T07:34:15Z"},{"alias_kind":"arxiv_version","alias_value":"2306.13649v3","created_at":"2026-07-05T07:34:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2306.13649","created_at":"2026-07-05T07:34:15Z"},{"alias_kind":"pith_short_12","alias_value":"3Z6ZFPYJIGTZ","created_at":"2026-07-05T07:34:15Z"},{"alias_kind":"pith_short_16","alias_value":"3Z6ZFPYJIGTZL23J","created_at":"2026-07-05T07:34:15Z"},{"alias_kind":"pith_short_8","alias_value":"3Z6ZFPYJ","created_at":"2026-07-05T07:34:15Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:3Z6ZFPYJIGTZL23JGJIZYXC6DR","target":"record","payload":{"canonical_record":{"source":{"id":"2306.13649","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-06-23T17:56:26Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"a1e839e0ebb01d6ed89cbf56a10842de6baaa7a05209b87ef5f411622e0fe35d","abstract_canon_sha256":"389ee6658f57e9b1a62394a4d26a57ea927da59024257f05bf61fe567e9be085"},"schema_version":"1.0"},"canonical_sha256":"de7d92bf0941a795eb6932519c5c5e1c5d1e0d359bdb17e8871a4d3f559c3fb6","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:34:15.598853Z","signature_b64":"nuzDoT0ZMNGfXIUFoyddMOx6SIyC+sqTnu0dMdSPmwalbmxW4ALJnQuYzS0za1nb4xp7x6yhigv8lbT8UTRLDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"de7d92bf0941a795eb6932519c5c5e1c5d1e0d359bdb17e8871a4d3f559c3fb6","last_reissued_at":"2026-07-05T07:34:15.598455Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:34:15.598455Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2306.13649","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T07:34:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"RP0Qz7C+oEORsIK9YtJnRlGFeWvP9iRGn8KB/8MDaMXgW9h3bya+zOK6VTKn+erF+yx+P2DIH4TjO+11SF7wCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-15T19:45:32.286688Z"},"content_sha256":"9169b8611c1186f9c684314f9979b098fbc7e585af047164efe2e1097db46ed8","schema_version":"1.0","event_id":"sha256:9169b8611c1186f9c684314f9979b098fbc7e585af047164efe2e1097db46ed8"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:3Z6ZFPYJIGTZL23JGJIZYXC6DR","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"On-Policy Distillation of Language Models: Learning from Self-Generated Mistakes","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Matthieu Geist, Nino Vieillard, Olivier Bachem, Piotr Stanczyk, Rishabh Agarwal, Sabela Ramos, Yongchao Zhou","submitted_at":"2023-06-23T17:56:26Z","abstract_excerpt":"Knowledge distillation (KD) is widely used for compressing a teacher model to reduce its inference cost and memory footprint, by training a smaller student model. However, current KD methods for auto-regressive sequence models suffer from distribution mismatch between output sequences seen during training and those generated by the student during inference. To address this issue, we introduce Generalized Knowledge Distillation (GKD). Instead of solely relying on a fixed set of output sequences, GKD trains the student on its self-generated output sequences by leveraging feedback from the teache"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2306.13649","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/2306.13649/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T07:34:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"OykUpC14fnDJvTTiKyZGySe3CQDHfTcSl6uh+O4jmga08nhSUxJAM2pJZAQrWLO128UfdtoJpYLfykMMy7E8BQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-15T19:45:32.287192Z"},"content_sha256":"709a9f78b7e14eceaa7dc9d7a10c437ea560445835990f5bcd67bdaae510da90","schema_version":"1.0","event_id":"sha256:709a9f78b7e14eceaa7dc9d7a10c437ea560445835990f5bcd67bdaae510da90"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/3Z6ZFPYJIGTZL23JGJIZYXC6DR/bundle.json","state_url":"https://pith.science/pith/3Z6ZFPYJIGTZL23JGJIZYXC6DR/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/3Z6ZFPYJIGTZL23JGJIZYXC6DR/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":"2026-08-15T19:45:32Z","links":{"resolver":"https://pith.science/pith/3Z6ZFPYJIGTZL23JGJIZYXC6DR","bundle":"https://pith.science/pith/3Z6ZFPYJIGTZL23JGJIZYXC6DR/bundle.json","state":"https://pith.science/pith/3Z6ZFPYJIGTZL23JGJIZYXC6DR/state.json","well_known_bundle":"https://pith.science/.well-known/pith/3Z6ZFPYJIGTZL23JGJIZYXC6DR/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:3Z6ZFPYJIGTZL23JGJIZYXC6DR","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":"389ee6658f57e9b1a62394a4d26a57ea927da59024257f05bf61fe567e9be085","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-06-23T17:56:26Z","title_canon_sha256":"a1e839e0ebb01d6ed89cbf56a10842de6baaa7a05209b87ef5f411622e0fe35d"},"schema_version":"1.0","source":{"id":"2306.13649","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2306.13649","created_at":"2026-07-05T07:34:15Z"},{"alias_kind":"arxiv_version","alias_value":"2306.13649v3","created_at":"2026-07-05T07:34:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2306.13649","created_at":"2026-07-05T07:34:15Z"},{"alias_kind":"pith_short_12","alias_value":"3Z6ZFPYJIGTZ","created_at":"2026-07-05T07:34:15Z"},{"alias_kind":"pith_short_16","alias_value":"3Z6ZFPYJIGTZL23J","created_at":"2026-07-05T07:34:15Z"},{"alias_kind":"pith_short_8","alias_value":"3Z6ZFPYJ","created_at":"2026-07-05T07:34:15Z"}],"graph_snapshots":[{"event_id":"sha256:709a9f78b7e14eceaa7dc9d7a10c437ea560445835990f5bcd67bdaae510da90","target":"graph","created_at":"2026-07-05T07:34:15Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2306.13649/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Knowledge distillation (KD) is widely used for compressing a teacher model to reduce its inference cost and memory footprint, by training a smaller student model. However, current KD methods for auto-regressive sequence models suffer from distribution mismatch between output sequences seen during training and those generated by the student during inference. To address this issue, we introduce Generalized Knowledge Distillation (GKD). Instead of solely relying on a fixed set of output sequences, GKD trains the student on its self-generated output sequences by leveraging feedback from the teache","authors_text":"Matthieu Geist, Nino Vieillard, Olivier Bachem, Piotr Stanczyk, Rishabh Agarwal, Sabela Ramos, Yongchao Zhou","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-06-23T17:56:26Z","title":"On-Policy Distillation of Language Models: Learning from Self-Generated Mistakes"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2306.13649","kind":"arxiv","version":3},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:9169b8611c1186f9c684314f9979b098fbc7e585af047164efe2e1097db46ed8","target":"record","created_at":"2026-07-05T07:34:15Z","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":"389ee6658f57e9b1a62394a4d26a57ea927da59024257f05bf61fe567e9be085","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-06-23T17:56:26Z","title_canon_sha256":"a1e839e0ebb01d6ed89cbf56a10842de6baaa7a05209b87ef5f411622e0fe35d"},"schema_version":"1.0","source":{"id":"2306.13649","kind":"arxiv","version":3}},"canonical_sha256":"de7d92bf0941a795eb6932519c5c5e1c5d1e0d359bdb17e8871a4d3f559c3fb6","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"de7d92bf0941a795eb6932519c5c5e1c5d1e0d359bdb17e8871a4d3f559c3fb6","first_computed_at":"2026-07-05T07:34:15.598455Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:34:15.598455Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"nuzDoT0ZMNGfXIUFoyddMOx6SIyC+sqTnu0dMdSPmwalbmxW4ALJnQuYzS0za1nb4xp7x6yhigv8lbT8UTRLDw==","signature_status":"signed_v1","signed_at":"2026-07-05T07:34:15.598853Z","signed_message":"canonical_sha256_bytes"},"source_id":"2306.13649","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9169b8611c1186f9c684314f9979b098fbc7e585af047164efe2e1097db46ed8","sha256:709a9f78b7e14eceaa7dc9d7a10c437ea560445835990f5bcd67bdaae510da90"],"state_sha256":"ca05b6124b5fe83db74d7253886e0ad0812d80bf322e6db2030d6bd912c1bd21"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"wGXIRZTE8os/HXncqLjstYTT3kCViAuEdbtzITLbPejrMgf0YvkLDahLp/eL+BfK34db5ldxefGlfPJA+QixAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-15T19:45:32.291586Z","bundle_sha256":"5d9fc28944b7c5b212cc1688671a5579eaf234a29db5848b32672f44797b49f4"}}