{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:MEVPW3SFTWGBNIEDJWW6LKPTJV","short_pith_number":"pith:MEVPW3SF","canonical_record":{"source":{"id":"2607.04728","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-07-06T06:59:44Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"b4f7be4d12e6a736be352ba363f1128d33e1c9c42e7f2ee747c76c4b55341337","abstract_canon_sha256":"03a3bccd8dd7ac1f560cbe6982995eb9b79c95d4284724e4696bb10c18c5680c"},"schema_version":"1.0"},"canonical_sha256":"612afb6e459d8c16a0834dade5a9f34d47cb5c26ef1aa8bf364334fd12eb7b0c","source":{"kind":"arxiv","id":"2607.04728","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.04728","created_at":"2026-07-07T02:20:00Z"},{"alias_kind":"arxiv_version","alias_value":"2607.04728v1","created_at":"2026-07-07T02:20:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.04728","created_at":"2026-07-07T02:20:00Z"},{"alias_kind":"pith_short_12","alias_value":"MEVPW3SFTWGB","created_at":"2026-07-07T02:20:00Z"},{"alias_kind":"pith_short_16","alias_value":"MEVPW3SFTWGBNIED","created_at":"2026-07-07T02:20:00Z"},{"alias_kind":"pith_short_8","alias_value":"MEVPW3SF","created_at":"2026-07-07T02:20:00Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:MEVPW3SFTWGBNIEDJWW6LKPTJV","target":"record","payload":{"canonical_record":{"source":{"id":"2607.04728","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-07-06T06:59:44Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"b4f7be4d12e6a736be352ba363f1128d33e1c9c42e7f2ee747c76c4b55341337","abstract_canon_sha256":"03a3bccd8dd7ac1f560cbe6982995eb9b79c95d4284724e4696bb10c18c5680c"},"schema_version":"1.0"},"canonical_sha256":"612afb6e459d8c16a0834dade5a9f34d47cb5c26ef1aa8bf364334fd12eb7b0c","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T02:20:00.167159Z","signature_b64":"08jQt4HPr8ftO4Sn2oqwLjo4YCO1g2CprHGbjt9L4AovzY41Uvuw7q0S3kwRPGcNj2aAOwlA7igmAPPnDohyCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"612afb6e459d8c16a0834dade5a9f34d47cb5c26ef1aa8bf364334fd12eb7b0c","last_reissued_at":"2026-07-07T02:20:00.166406Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T02:20:00.166406Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2607.04728","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":"2026-07-07T02:20:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"kfIBQfGbeYuODjFH0pgD0dO0WcSTTUvi7nzKhD+gor8j45GCnQ15Q7ukHG9RTfD+8hq3eryMv1hXVDpk5hhHBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T15:32:41.760771Z"},"content_sha256":"3f20947b8db0d3c887c9c268ea7550ae4caaa2a72db7debd4a38468b73a097af","schema_version":"1.0","event_id":"sha256:3f20947b8db0d3c887c9c268ea7550ae4caaa2a72db7debd4a38468b73a097af"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:MEVPW3SFTWGBNIEDJWW6LKPTJV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Turning Off-Policy Tokens On-Policy: A Plug-in Approach for Improving LLM Alignment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Jiaxing Wang, Mingyang Yi, Xiuyu Li, Yu Li, zhangliangxu, Zhaolong Xing, Zhen Chen","submitted_at":"2026-07-06T06:59:44Z","abstract_excerpt":"Reinforcement learning (RL) post-training for large language models (LLMs) follows a efficient paradigm of \"rollout then update\", which inevitably results in off-policy training data. To resolve this, Importance sampling (IS) is proposed, while the token-level ratios compound over long sequences, causing severe variance exploded. A natural idea is \"transferring\" these off-policy token into on-policy token, so that the importance scores for correction are unnecessary. Following this idea, we propose Selective Importance Sampling (SIS), which is inspired by rejection sampling. Concretely, SIS im"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.04728","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/2607.04728/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":"2026-07-07T02:20:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"NXU3Tzuc/oYrmS0sSVAh+2//dcfQmjmPs9hJyyiHZoK7WDRSBBYmvnfkh6Zu/UW1tPcvVg2hjemMYhFdbiLdDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T15:32:41.761309Z"},"content_sha256":"1452db7b9f3b852117a1b5b15e73c36401ca1c339768a5fa3d5386b208c61225","schema_version":"1.0","event_id":"sha256:1452db7b9f3b852117a1b5b15e73c36401ca1c339768a5fa3d5386b208c61225"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/MEVPW3SFTWGBNIEDJWW6LKPTJV/bundle.json","state_url":"https://pith.science/pith/MEVPW3SFTWGBNIEDJWW6LKPTJV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/MEVPW3SFTWGBNIEDJWW6LKPTJV/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":"2026-08-21T15:32:41Z","links":{"resolver":"https://pith.science/pith/MEVPW3SFTWGBNIEDJWW6LKPTJV","bundle":"https://pith.science/pith/MEVPW3SFTWGBNIEDJWW6LKPTJV/bundle.json","state":"https://pith.science/pith/MEVPW3SFTWGBNIEDJWW6LKPTJV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/MEVPW3SFTWGBNIEDJWW6LKPTJV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:MEVPW3SFTWGBNIEDJWW6LKPTJV","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":"03a3bccd8dd7ac1f560cbe6982995eb9b79c95d4284724e4696bb10c18c5680c","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-07-06T06:59:44Z","title_canon_sha256":"b4f7be4d12e6a736be352ba363f1128d33e1c9c42e7f2ee747c76c4b55341337"},"schema_version":"1.0","source":{"id":"2607.04728","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.04728","created_at":"2026-07-07T02:20:00Z"},{"alias_kind":"arxiv_version","alias_value":"2607.04728v1","created_at":"2026-07-07T02:20:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.04728","created_at":"2026-07-07T02:20:00Z"},{"alias_kind":"pith_short_12","alias_value":"MEVPW3SFTWGB","created_at":"2026-07-07T02:20:00Z"},{"alias_kind":"pith_short_16","alias_value":"MEVPW3SFTWGBNIED","created_at":"2026-07-07T02:20:00Z"},{"alias_kind":"pith_short_8","alias_value":"MEVPW3SF","created_at":"2026-07-07T02:20:00Z"}],"graph_snapshots":[{"event_id":"sha256:1452db7b9f3b852117a1b5b15e73c36401ca1c339768a5fa3d5386b208c61225","target":"graph","created_at":"2026-07-07T02:20:00Z","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/2607.04728/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reinforcement learning (RL) post-training for large language models (LLMs) follows a efficient paradigm of \"rollout then update\", which inevitably results in off-policy training data. To resolve this, Importance sampling (IS) is proposed, while the token-level ratios compound over long sequences, causing severe variance exploded. A natural idea is \"transferring\" these off-policy token into on-policy token, so that the importance scores for correction are unnecessary. Following this idea, we propose Selective Importance Sampling (SIS), which is inspired by rejection sampling. Concretely, SIS im","authors_text":"Jiaxing Wang, Mingyang Yi, Xiuyu Li, Yu Li, zhangliangxu, Zhaolong Xing, Zhen Chen","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-07-06T06:59:44Z","title":"Turning Off-Policy Tokens On-Policy: A Plug-in Approach for Improving LLM Alignment"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.04728","kind":"arxiv","version":1},"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":"sha256:3f20947b8db0d3c887c9c268ea7550ae4caaa2a72db7debd4a38468b73a097af","target":"record","created_at":"2026-07-07T02:20:00Z","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":"03a3bccd8dd7ac1f560cbe6982995eb9b79c95d4284724e4696bb10c18c5680c","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-07-06T06:59:44Z","title_canon_sha256":"b4f7be4d12e6a736be352ba363f1128d33e1c9c42e7f2ee747c76c4b55341337"},"schema_version":"1.0","source":{"id":"2607.04728","kind":"arxiv","version":1}},"canonical_sha256":"612afb6e459d8c16a0834dade5a9f34d47cb5c26ef1aa8bf364334fd12eb7b0c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"612afb6e459d8c16a0834dade5a9f34d47cb5c26ef1aa8bf364334fd12eb7b0c","first_computed_at":"2026-07-07T02:20:00.166406Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-07T02:20:00.166406Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"08jQt4HPr8ftO4Sn2oqwLjo4YCO1g2CprHGbjt9L4AovzY41Uvuw7q0S3kwRPGcNj2aAOwlA7igmAPPnDohyCQ==","signature_status":"signed_v1","signed_at":"2026-07-07T02:20:00.167159Z","signed_message":"canonical_sha256_bytes"},"source_id":"2607.04728","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3f20947b8db0d3c887c9c268ea7550ae4caaa2a72db7debd4a38468b73a097af","sha256:1452db7b9f3b852117a1b5b15e73c36401ca1c339768a5fa3d5386b208c61225"],"state_sha256":"4ece40d2a4fd1eef7a90516384d85fca1811ef8f114a8f190b44cf7921b879ae"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1ERPudbIX542MpT3ZICzGP9OAyVK4O43i/g1BiqjvI2521/uMv9MbLRsLbRVk4OeW76ha4ZXGjplH4fPukD4BQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-21T15:32:41.766148Z","bundle_sha256":"3894f8106b8e47e10380429d96b0b0330cb5e5d14727163d1417c175bbe17a52"}}