{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:W2BPVN54DJJ4MBSGOPBNYGGSF7","short_pith_number":"pith:W2BPVN54","canonical_record":{"source":{"id":"2602.10346","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-02-10T22:36:48Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"207a25f7eb6e7a84f5dbf714060145fa53a1cf5ee3a311dc9132c9625550e51f","abstract_canon_sha256":"93bad57a54f8ed58cfe362360bef037fc26063aa1c94a6c8fcd627c5bb285320"},"schema_version":"1.0"},"canonical_sha256":"b682fab7bc1a53c6064673c2dc18d22ff1b8f33c5a27b44fd4f5f25b0dc6a5c1","source":{"kind":"arxiv","id":"2602.10346","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.10346","created_at":"2026-05-17T23:39:16Z"},{"alias_kind":"arxiv_version","alias_value":"2602.10346v2","created_at":"2026-05-17T23:39:16Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.10346","created_at":"2026-05-17T23:39:16Z"},{"alias_kind":"pith_short_12","alias_value":"W2BPVN54DJJ4","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"W2BPVN54DJJ4MBSG","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"W2BPVN54","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:W2BPVN54DJJ4MBSGOPBNYGGSF7","target":"record","payload":{"canonical_record":{"source":{"id":"2602.10346","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-02-10T22:36:48Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"207a25f7eb6e7a84f5dbf714060145fa53a1cf5ee3a311dc9132c9625550e51f","abstract_canon_sha256":"93bad57a54f8ed58cfe362360bef037fc26063aa1c94a6c8fcd627c5bb285320"},"schema_version":"1.0"},"canonical_sha256":"b682fab7bc1a53c6064673c2dc18d22ff1b8f33c5a27b44fd4f5f25b0dc6a5c1","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:39:16.223443Z","signature_b64":"nn2YzvMnLUgJW0sopn0ZM4AbWfWWPG2o3SxjYUYRxRm6Dc3ihpYQz8JiQJE8J6nUIqLKKrPo+nEIJfGk+7tiAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b682fab7bc1a53c6064673c2dc18d22ff1b8f33c5a27b44fd4f5f25b0dc6a5c1","last_reissued_at":"2026-05-17T23:39:16.222671Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:39:16.222671Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2602.10346","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":"2026-05-17T23:39:16Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"YGmL0WvVdbzGiXsCEpOcPLxTA2nZVectOCZMSoBwlZSp0igeIT6Qy8v/XIT7xhAtUH8QvC0f8Vv/uT5ANWv5Cw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-28T12:16:18.729967Z"},"content_sha256":"ac429a6711653d3e3b514b0cab16235e168dc440c17074fab8460191446ac65e","schema_version":"1.0","event_id":"sha256:ac429a6711653d3e3b514b0cab16235e168dc440c17074fab8460191446ac65e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:W2BPVN54DJJ4MBSGOPBNYGGSF7","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Geometry-Aware Decoding with Wasserstein-Regularized Truncation and Mass Penalties for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Wasserstein distance over token embeddings yields a closed-form truncation rule that balances mass and entropy in LLM decoding.","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Arash Gholami Davoodi, Navid Rezazadeh, Pouya Pezeshkpour, Seyed Pouyan Mousavi Davoudi","submitted_at":"2026-02-10T22:36:48Z","abstract_excerpt":"Large language models (LLMs) must balance diversity and creativity against logical coherence in open-ended generation. Existing truncation-based samplers are effective but largely heuristic, relying mainly on probability mass and entropy while ignoring semantic geometry of the token space. We present Top-W, a geometry-aware truncation rule that uses Wasserstein distance-defined over token-embedding geometry-to keep the cropped distribution close to the original, while explicitly balancing retained probability mass against the entropy of the kept set. Our theory yields a simple closed-form stru"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our theory yields a simple closed-form structure for the fixed-potential subset update... achieving up to 33.7% improvement.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That Wasserstein distance defined over token-embedding geometry meaningfully captures semantic relationships required for logical coherence during open-ended generation.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Top-W applies Wasserstein-regularized truncation on token-embedding geometry to create a closed-form optimal crop for LLM sampling that outperforms prior methods by up to 33.7% on GSM8K, GPQA, AlpacaEval, and MT-Bench.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Wasserstein distance over token embeddings yields a closed-form truncation rule that balances mass and entropy in LLM decoding.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"1adc735ef44b66f27ea7aaec6156e20a3a54abd46bc97bd55d16eac574e88f77"},"source":{"id":"2602.10346","kind":"arxiv","version":2},"verdict":{"id":"32b662c8-6905-47ec-afd7-966a0301f4c3","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T05:05:01.174568Z","strongest_claim":"Our theory yields a simple closed-form structure for the fixed-potential subset update... achieving up to 33.7% improvement.","one_line_summary":"Top-W applies Wasserstein-regularized truncation on token-embedding geometry to create a closed-form optimal crop for LLM sampling that outperforms prior methods by up to 33.7% on GSM8K, GPQA, AlpacaEval, and MT-Bench.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That Wasserstein distance defined over token-embedding geometry meaningfully captures semantic relationships required for logical coherence during open-ended generation.","pith_extraction_headline":"Wasserstein distance over token embeddings yields a closed-form truncation rule that balances mass and entropy in LLM decoding."},"references":{"count":21,"sample":[{"doi":"","year":null,"title":"Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone","work_id":"feef9556-a016-493c-abd2-0c97a23a7ebf","ref_index":1,"cited_arxiv_id":"2404.14219","is_internal_anchor":true},{"doi":"","year":2007,"title":"Mirostat: A perplexity-controlled neural text decoding algorithm","work_id":"adf6dca4-9ef0-4476-8003-87f72f5b99c9","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1901,"title":"Lan- guage models are few-shot learners.Advances in neural information processing systems, 33:1877–1901,","work_id":"ed7e2043-f7e9-45d2-8c60-61778195b799","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"DoLa: Decoding by Contrasting Layers Improves Factuality in Large Language Models","work_id":"0f1b9a7a-0623-4efc-bed3-6dd997054681","ref_index":4,"cited_arxiv_id":"2309.03883","is_internal_anchor":true},{"doi":"","year":null,"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","ref_index":5,"cited_arxiv_id":"2110.14168","is_internal_anchor":true}],"resolved_work":21,"snapshot_sha256":"2efd99269372ebe9905f81d5eb95d97ec9cd1bb7285edd28853f5e623163a823","internal_anchors":12},"formal_canon":{"evidence_count":2,"snapshot_sha256":"ea470f9d7b1f7e895f2961c66cae3348c4867833af2643ae7ad6e74bbb2cddb7"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"32b662c8-6905-47ec-afd7-966a0301f4c3"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:39:16Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"FcutEBycizU2lxdwvTxWVnq7XtFGqzKGIUkgngTBU3qgycVg3ycDRWObiAj8rNZtXnzMWOQRpi2HTyB7HD0NAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-28T12:16:18.730601Z"},"content_sha256":"27c409b9a7eee1dfa82de7d5ed97733582108964a55c685c3323f83e99a6e40b","schema_version":"1.0","event_id":"sha256:27c409b9a7eee1dfa82de7d5ed97733582108964a55c685c3323f83e99a6e40b"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/W2BPVN54DJJ4MBSGOPBNYGGSF7/bundle.json","state_url":"https://pith.science/pith/W2BPVN54DJJ4MBSGOPBNYGGSF7/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/W2BPVN54DJJ4MBSGOPBNYGGSF7/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":"2026-05-28T12:16:18Z","links":{"resolver":"https://pith.science/pith/W2BPVN54DJJ4MBSGOPBNYGGSF7","bundle":"https://pith.science/pith/W2BPVN54DJJ4MBSGOPBNYGGSF7/bundle.json","state":"https://pith.science/pith/W2BPVN54DJJ4MBSGOPBNYGGSF7/state.json","well_known_bundle":"https://pith.science/.well-known/pith/W2BPVN54DJJ4MBSGOPBNYGGSF7/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:W2BPVN54DJJ4MBSGOPBNYGGSF7","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":"93bad57a54f8ed58cfe362360bef037fc26063aa1c94a6c8fcd627c5bb285320","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-02-10T22:36:48Z","title_canon_sha256":"207a25f7eb6e7a84f5dbf714060145fa53a1cf5ee3a311dc9132c9625550e51f"},"schema_version":"1.0","source":{"id":"2602.10346","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.10346","created_at":"2026-05-17T23:39:16Z"},{"alias_kind":"arxiv_version","alias_value":"2602.10346v2","created_at":"2026-05-17T23:39:16Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.10346","created_at":"2026-05-17T23:39:16Z"},{"alias_kind":"pith_short_12","alias_value":"W2BPVN54DJJ4","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"W2BPVN54DJJ4MBSG","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"W2BPVN54","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:27c409b9a7eee1dfa82de7d5ed97733582108964a55c685c3323f83e99a6e40b","target":"graph","created_at":"2026-05-17T23:39:16Z","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":"Our theory yields a simple closed-form structure for the fixed-potential subset update... achieving up to 33.7% improvement."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That Wasserstein distance defined over token-embedding geometry meaningfully captures semantic relationships required for logical coherence during open-ended generation."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Top-W applies Wasserstein-regularized truncation on token-embedding geometry to create a closed-form optimal crop for LLM sampling that outperforms prior methods by up to 33.7% on GSM8K, GPQA, AlpacaEval, and MT-Bench."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Wasserstein distance over token embeddings yields a closed-form truncation rule that balances mass and entropy in LLM decoding."}],"snapshot_sha256":"1adc735ef44b66f27ea7aaec6156e20a3a54abd46bc97bd55d16eac574e88f77"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"ea470f9d7b1f7e895f2961c66cae3348c4867833af2643ae7ad6e74bbb2cddb7"},"paper":{"abstract_excerpt":"Large language models (LLMs) must balance diversity and creativity against logical coherence in open-ended generation. Existing truncation-based samplers are effective but largely heuristic, relying mainly on probability mass and entropy while ignoring semantic geometry of the token space. We present Top-W, a geometry-aware truncation rule that uses Wasserstein distance-defined over token-embedding geometry-to keep the cropped distribution close to the original, while explicitly balancing retained probability mass against the entropy of the kept set. Our theory yields a simple closed-form stru","authors_text":"Arash Gholami Davoodi, Navid Rezazadeh, Pouya Pezeshkpour, Seyed Pouyan Mousavi Davoudi","cross_cats":["cs.LG"],"headline":"Wasserstein distance over token embeddings yields a closed-form truncation rule that balances mass and entropy in LLM decoding.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-02-10T22:36:48Z","title":"Geometry-Aware Decoding with Wasserstein-Regularized Truncation and Mass Penalties for Large Language Models"},"references":{"count":21,"internal_anchors":12,"resolved_work":21,"sample":[{"cited_arxiv_id":"2404.14219","doi":"","is_internal_anchor":true,"ref_index":1,"title":"Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone","work_id":"feef9556-a016-493c-abd2-0c97a23a7ebf","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Mirostat: A perplexity-controlled neural text decoding algorithm","work_id":"adf6dca4-9ef0-4476-8003-87f72f5b99c9","year":2007},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Lan- guage models are few-shot learners.Advances in neural information processing systems, 33:1877–1901,","work_id":"ed7e2043-f7e9-45d2-8c60-61778195b799","year":1901},{"cited_arxiv_id":"2309.03883","doi":"","is_internal_anchor":true,"ref_index":4,"title":"DoLa: Decoding by Contrasting Layers Improves Factuality in Large Language Models","work_id":"0f1b9a7a-0623-4efc-bed3-6dd997054681","year":null},{"cited_arxiv_id":"2110.14168","doi":"","is_internal_anchor":true,"ref_index":5,"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","year":null}],"snapshot_sha256":"2efd99269372ebe9905f81d5eb95d97ec9cd1bb7285edd28853f5e623163a823"},"source":{"id":"2602.10346","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-16T05:05:01.174568Z","id":"32b662c8-6905-47ec-afd7-966a0301f4c3","model_set":{"reader":"grok-4.3"},"one_line_summary":"Top-W applies Wasserstein-regularized truncation on token-embedding geometry to create a closed-form optimal crop for LLM sampling that outperforms prior methods by up to 33.7% on GSM8K, GPQA, AlpacaEval, and MT-Bench.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Wasserstein distance over token embeddings yields a closed-form truncation rule that balances mass and entropy in LLM decoding.","strongest_claim":"Our theory yields a simple closed-form structure for the fixed-potential subset update... achieving up to 33.7% improvement.","weakest_assumption":"That Wasserstein distance defined over token-embedding geometry meaningfully captures semantic relationships required for logical coherence during open-ended generation."}},"verdict_id":"32b662c8-6905-47ec-afd7-966a0301f4c3"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:ac429a6711653d3e3b514b0cab16235e168dc440c17074fab8460191446ac65e","target":"record","created_at":"2026-05-17T23:39:16Z","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":"93bad57a54f8ed58cfe362360bef037fc26063aa1c94a6c8fcd627c5bb285320","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-02-10T22:36:48Z","title_canon_sha256":"207a25f7eb6e7a84f5dbf714060145fa53a1cf5ee3a311dc9132c9625550e51f"},"schema_version":"1.0","source":{"id":"2602.10346","kind":"arxiv","version":2}},"canonical_sha256":"b682fab7bc1a53c6064673c2dc18d22ff1b8f33c5a27b44fd4f5f25b0dc6a5c1","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"b682fab7bc1a53c6064673c2dc18d22ff1b8f33c5a27b44fd4f5f25b0dc6a5c1","first_computed_at":"2026-05-17T23:39:16.222671Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:39:16.222671Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"nn2YzvMnLUgJW0sopn0ZM4AbWfWWPG2o3SxjYUYRxRm6Dc3ihpYQz8JiQJE8J6nUIqLKKrPo+nEIJfGk+7tiAg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:39:16.223443Z","signed_message":"canonical_sha256_bytes"},"source_id":"2602.10346","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:ac429a6711653d3e3b514b0cab16235e168dc440c17074fab8460191446ac65e","sha256:27c409b9a7eee1dfa82de7d5ed97733582108964a55c685c3323f83e99a6e40b"],"state_sha256":"dd9a47be0744ee8561ee1aeefdea7c15b375251b428546ced3315560fa9273d4"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Ak9I+yutehhC6daznihPQS29CMI2W1NQVNbfz2P+JaE2Wp4H6Vy5aXtnR1cr+4gosR5ZTpsyZRlYz8bQquFCBg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-28T12:16:18.733216Z","bundle_sha256":"8ac73650c12f4076f0a50aa9c8a439a6f2f6bf2f2d032e8e62bb574caa61771b"}}