{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:JCSWQ43WCHMSDLIN7AZ4QRQ4D3","short_pith_number":"pith:JCSWQ43W","canonical_record":{"source":{"id":"2410.14655","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-18T17:48:27Z","cross_cats_sorted":[],"title_canon_sha256":"c8f3d8c11294c5c538438d6e1085737fd8daf1f712d9bafa7402c873511dd698","abstract_canon_sha256":"8e2cef2858c12c541632956f6beef71d7e7c525ae3f482be0c2453ba0e7f5f73"},"schema_version":"1.0"},"canonical_sha256":"48a568737611d921ad0df833c8461c1ef75fcc53d36f47f7e5b8572f303da3a8","source":{"kind":"arxiv","id":"2410.14655","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.14655","created_at":"2026-07-05T10:03:28Z"},{"alias_kind":"arxiv_version","alias_value":"2410.14655v2","created_at":"2026-07-05T10:03:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.14655","created_at":"2026-07-05T10:03:28Z"},{"alias_kind":"pith_short_12","alias_value":"JCSWQ43WCHMS","created_at":"2026-07-05T10:03:28Z"},{"alias_kind":"pith_short_16","alias_value":"JCSWQ43WCHMSDLIN","created_at":"2026-07-05T10:03:28Z"},{"alias_kind":"pith_short_8","alias_value":"JCSWQ43W","created_at":"2026-07-05T10:03:28Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:JCSWQ43WCHMSDLIN7AZ4QRQ4D3","target":"record","payload":{"canonical_record":{"source":{"id":"2410.14655","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-18T17:48:27Z","cross_cats_sorted":[],"title_canon_sha256":"c8f3d8c11294c5c538438d6e1085737fd8daf1f712d9bafa7402c873511dd698","abstract_canon_sha256":"8e2cef2858c12c541632956f6beef71d7e7c525ae3f482be0c2453ba0e7f5f73"},"schema_version":"1.0"},"canonical_sha256":"48a568737611d921ad0df833c8461c1ef75fcc53d36f47f7e5b8572f303da3a8","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:03:28.226532Z","signature_b64":"g3QlT+N6c7vABc/g0OrnywaZ1f0hWPMXKXSvwNhSnPLtS+M5aWvlr6fc8yWLJk/Q8t9Cqz7YJXzIzWNVjgiWAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"48a568737611d921ad0df833c8461c1ef75fcc53d36f47f7e5b8572f303da3a8","last_reissued_at":"2026-07-05T10:03:28.226082Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:03:28.226082Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.14655","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:03:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Fc6eNiRQNW5Bmyx7NMAA5AQa0AyYcY0TO2AFy6iyTf6tUC4m0RqSaBqcju8eSBS5FFFdkaMl+AYKjH8dxn7sBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T11:12:59.726216Z"},"content_sha256":"63a4c775da1ac62ab5884addd5c1742c44454a8cef048a48a898b9f3fe461f0e","schema_version":"1.0","event_id":"sha256:63a4c775da1ac62ab5884addd5c1742c44454a8cef048a48a898b9f3fe461f0e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:JCSWQ43WCHMSDLIN7AZ4QRQ4D3","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Bridging the Training-Inference Gap in LLMs by Leveraging Self-Generated Tokens","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"George Karypis, Huzefa Rangwala, Pratik Chaudhari, Rasool Fakoor, Siliang Zeng, Yao Liu, Zhepeng Cen","submitted_at":"2024-10-18T17:48:27Z","abstract_excerpt":"Language models are often trained to maximize the likelihood of the next token given past tokens in the training dataset. However, during inference time, they are utilized differently, generating text sequentially and auto-regressively by using previously generated tokens as input to predict the next one. Marginal differences in predictions at each step can cascade over successive steps, resulting in different distributions from what the models were trained for and potentially leading to unpredictable behavior. This paper proposes two simple approaches based on model own generation to address "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.14655","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/2410.14655/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":"2026-07-05T10:03:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Fn3mUQowzTItj3afPNUmvwpohACahj4bO4sSgaGz+EtjfPvf/AFUx2YVUAoXb+mXVABRlMsogSfx0YyKZZTnBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T11:12:59.726566Z"},"content_sha256":"2e5afa9c468230338ba4bd44a34600d1c0ae57e66da84aab5ca98aca5900c481","schema_version":"1.0","event_id":"sha256:2e5afa9c468230338ba4bd44a34600d1c0ae57e66da84aab5ca98aca5900c481"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/JCSWQ43WCHMSDLIN7AZ4QRQ4D3/bundle.json","state_url":"https://pith.science/pith/JCSWQ43WCHMSDLIN7AZ4QRQ4D3/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/JCSWQ43WCHMSDLIN7AZ4QRQ4D3/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":"2026-08-05T11:12:59Z","links":{"resolver":"https://pith.science/pith/JCSWQ43WCHMSDLIN7AZ4QRQ4D3","bundle":"https://pith.science/pith/JCSWQ43WCHMSDLIN7AZ4QRQ4D3/bundle.json","state":"https://pith.science/pith/JCSWQ43WCHMSDLIN7AZ4QRQ4D3/state.json","well_known_bundle":"https://pith.science/.well-known/pith/JCSWQ43WCHMSDLIN7AZ4QRQ4D3/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:JCSWQ43WCHMSDLIN7AZ4QRQ4D3","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":"8e2cef2858c12c541632956f6beef71d7e7c525ae3f482be0c2453ba0e7f5f73","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-18T17:48:27Z","title_canon_sha256":"c8f3d8c11294c5c538438d6e1085737fd8daf1f712d9bafa7402c873511dd698"},"schema_version":"1.0","source":{"id":"2410.14655","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.14655","created_at":"2026-07-05T10:03:28Z"},{"alias_kind":"arxiv_version","alias_value":"2410.14655v2","created_at":"2026-07-05T10:03:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.14655","created_at":"2026-07-05T10:03:28Z"},{"alias_kind":"pith_short_12","alias_value":"JCSWQ43WCHMS","created_at":"2026-07-05T10:03:28Z"},{"alias_kind":"pith_short_16","alias_value":"JCSWQ43WCHMSDLIN","created_at":"2026-07-05T10:03:28Z"},{"alias_kind":"pith_short_8","alias_value":"JCSWQ43W","created_at":"2026-07-05T10:03:28Z"}],"graph_snapshots":[{"event_id":"sha256:2e5afa9c468230338ba4bd44a34600d1c0ae57e66da84aab5ca98aca5900c481","target":"graph","created_at":"2026-07-05T10:03:28Z","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/2410.14655/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Language models are often trained to maximize the likelihood of the next token given past tokens in the training dataset. However, during inference time, they are utilized differently, generating text sequentially and auto-regressively by using previously generated tokens as input to predict the next one. Marginal differences in predictions at each step can cascade over successive steps, resulting in different distributions from what the models were trained for and potentially leading to unpredictable behavior. This paper proposes two simple approaches based on model own generation to address ","authors_text":"George Karypis, Huzefa Rangwala, Pratik Chaudhari, Rasool Fakoor, Siliang Zeng, Yao Liu, Zhepeng Cen","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-18T17:48:27Z","title":"Bridging the Training-Inference Gap in LLMs by Leveraging Self-Generated Tokens"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.14655","kind":"arxiv","version":2},"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":"sha256:63a4c775da1ac62ab5884addd5c1742c44454a8cef048a48a898b9f3fe461f0e","target":"record","created_at":"2026-07-05T10:03:28Z","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":"8e2cef2858c12c541632956f6beef71d7e7c525ae3f482be0c2453ba0e7f5f73","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-18T17:48:27Z","title_canon_sha256":"c8f3d8c11294c5c538438d6e1085737fd8daf1f712d9bafa7402c873511dd698"},"schema_version":"1.0","source":{"id":"2410.14655","kind":"arxiv","version":2}},"canonical_sha256":"48a568737611d921ad0df833c8461c1ef75fcc53d36f47f7e5b8572f303da3a8","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"48a568737611d921ad0df833c8461c1ef75fcc53d36f47f7e5b8572f303da3a8","first_computed_at":"2026-07-05T10:03:28.226082Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:03:28.226082Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"g3QlT+N6c7vABc/g0OrnywaZ1f0hWPMXKXSvwNhSnPLtS+M5aWvlr6fc8yWLJk/Q8t9Cqz7YJXzIzWNVjgiWAA==","signature_status":"signed_v1","signed_at":"2026-07-05T10:03:28.226532Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.14655","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:63a4c775da1ac62ab5884addd5c1742c44454a8cef048a48a898b9f3fe461f0e","sha256:2e5afa9c468230338ba4bd44a34600d1c0ae57e66da84aab5ca98aca5900c481"],"state_sha256":"60b942b446dbcb94b9b0c24e6bbd341afbebec55fef7e544c4b4536e517bcf71"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"SKFV187LSnap7N3wxotosrs+UpKAHDy9x8aZ1yYyFh5qi7/PEzZu6+Wj2m/nDjeHvz1YEjdc7Wngd2i+DlSmCA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-05T11:12:59.730222Z","bundle_sha256":"f1c2920dc961e7df7014e01cfd50d18758dd1a95b310fab180d1df1709376492"}}