{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:XZ6KU673RXJCDFKJNQGACJNLRV","short_pith_number":"pith:XZ6KU673","canonical_record":{"source":{"id":"2408.15664","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-08-28T09:31:09Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"2ba3f20877dadd522332bfbf50455cc66980575dd5bbc31e024cf536d29778ff","abstract_canon_sha256":"de6550001f65ea19f000f8d635df0f482bea715eade709c791379eef9256cdc5"},"schema_version":"1.0"},"canonical_sha256":"be7caa7bfb8dd22195496c0c0125ab8d718011f0f205abb2232fbcaf12a94cf6","source":{"kind":"arxiv","id":"2408.15664","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2408.15664","created_at":"2026-05-17T23:38:52Z"},{"alias_kind":"arxiv_version","alias_value":"2408.15664v1","created_at":"2026-05-17T23:38:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.15664","created_at":"2026-05-17T23:38:52Z"},{"alias_kind":"pith_short_12","alias_value":"XZ6KU673RXJC","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"XZ6KU673RXJCDFKJ","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"XZ6KU673","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:XZ6KU673RXJCDFKJNQGACJNLRV","target":"record","payload":{"canonical_record":{"source":{"id":"2408.15664","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-08-28T09:31:09Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"2ba3f20877dadd522332bfbf50455cc66980575dd5bbc31e024cf536d29778ff","abstract_canon_sha256":"de6550001f65ea19f000f8d635df0f482bea715eade709c791379eef9256cdc5"},"schema_version":"1.0"},"canonical_sha256":"be7caa7bfb8dd22195496c0c0125ab8d718011f0f205abb2232fbcaf12a94cf6","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:52.512176Z","signature_b64":"abHUXc0+nMOVci7mAwA8Ws1A0puaAk9b/wdZ9s1pIhqDSJV+b4lyfQpKD9+BIZ8Y4mimhyzkIYZThaggbRIrAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"be7caa7bfb8dd22195496c0c0125ab8d718011f0f205abb2232fbcaf12a94cf6","last_reissued_at":"2026-05-17T23:38:52.511754Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:52.511754Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2408.15664","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":"2026-05-17T23:38:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"FyO612efoqYpMFMQrqHf7iG4pDTR1G6JqCx0BMnTmk0MTIXHHrq89WW0VRF/5Ubvg+Bm0ikmAL6XVvPb33t+AQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-27T00:31:20.948028Z"},"content_sha256":"1762093ea14cc86b7cc32aeecf1835c4f30c04ba0667fbec5a5146f2fbcd5ff1","schema_version":"1.0","event_id":"sha256:1762093ea14cc86b7cc32aeecf1835c4f30c04ba0667fbec5a5146f2fbcd5ff1"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:XZ6KU673RXJCDFKJNQGACJNLRV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Auxiliary-Loss-Free Load Balancing Strategy for Mixture-of-Experts","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Mixture-of-Experts models reach higher performance with load balancing that avoids auxiliary loss gradients.","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Chenggang Zhao, Damai Dai, Huazuo Gao, Lean Wang, Xu Sun","submitted_at":"2024-08-28T09:31:09Z","abstract_excerpt":"For Mixture-of-Experts (MoE) models, an unbalanced expert load will lead to routing collapse or increased computational overhead. Existing methods commonly employ an auxiliary loss to encourage load balance, but a large auxiliary loss will introduce non-negligible interference gradients into training and thus impair the model performance. In order to control load balance while not producing undesired gradients during training, we propose Loss-Free Balancing, featured by an auxiliary-loss-free load balancing strategy. To be specific, before the top-K routing decision, Loss-Free Balancing will f"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Loss-Free Balancing achieves both better performance and better load balance compared with traditional auxiliary-loss-controlled load balancing strategies.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That dynamically updating per-expert biases from recent load statistics will maintain balance across training without introducing instability or unintended routing dynamics.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Loss-Free Balancing keeps expert loads balanced in MoE models by dynamically adjusting routing-score biases based on recent usage, avoiding auxiliary-loss interference and yielding better performance.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Mixture-of-Experts models reach higher performance with load balancing that avoids auxiliary loss gradients.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"c306008b14ccbd0298b01962067596270de956a35d5f098978a086983d20b104"},"source":{"id":"2408.15664","kind":"arxiv","version":1},"verdict":{"id":"9ed35d07-9958-4ddc-9949-761036dfa418","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T12:50:11.218526Z","strongest_claim":"Loss-Free Balancing achieves both better performance and better load balance compared with traditional auxiliary-loss-controlled load balancing strategies.","one_line_summary":"Loss-Free Balancing keeps expert loads balanced in MoE models by dynamically adjusting routing-score biases based on recent usage, avoiding auxiliary-loss interference and yielding better performance.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That dynamically updating per-expert biases from recent load statistics will maintain balance across training without introducing instability or unintended routing dynamics.","pith_extraction_headline":"Mixture-of-Experts models reach higher performance with load balancing that avoids auxiliary loss gradients."},"references":{"count":32,"sample":[{"doi":"","year":2024,"title":"DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models","work_id":"a9888d6d-bf47-4324-9834-7cc12ac3a78c","ref_index":1,"cited_arxiv_id":"2401.06066","is_internal_anchor":true},{"doi":"","year":2024,"title":"DeepSeek-Coder-V2: Breaking the Barrier of Closed-Source Models in Code Intelligence","work_id":"713a39be-8c4e-4ee3-8671-11cf9379262f","ref_index":2,"cited_arxiv_id":"2406.11931","is_internal_anchor":true},{"doi":"","year":2021,"title":"William Fedus, Barret Zoph, and Noam M. Shazeer. Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity. J. Mach. Learn. Res., 23: 0 120:1--120:39, 2021. URL http","work_id":"2f29f1ca-bd51-4381-acc7-95e946cde35f","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2006,"title":"GShard: Scaling Giant Models with Conditional Computation and Automatic Sharding","work_id":"52b3c9a6-2a27-45a7-ba2b-ebe4b5bb5a5f","ref_index":4,"cited_arxiv_id":"2006.16668","is_internal_anchor":true},{"doi":"","year":2016,"title":"Sgdr: Stochastic gradient descent with warm restarts","work_id":"21d4473d-9c67-4464-b3ea-96588e1d01e1","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":32,"snapshot_sha256":"45ecc5e36e60cebbf268394ae511828f42a1187c8166048c5f0e7f95a94d51e0","internal_anchors":6},"formal_canon":{"evidence_count":2,"snapshot_sha256":"ae8508b9f31a7ff7425541690bf1452fc90bde2e6779ab251d817b0d51cca3d1"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"9ed35d07-9958-4ddc-9949-761036dfa418"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ezSMSNADU2LLShKQjNnwL5gprDKLYCD2GqfZUk2DdnHB3zmnxYAdwl/Ky+EP/J6skDzE6elkOIbJIRLdst4jCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-27T00:31:20.949018Z"},"content_sha256":"2cf8798bc6ead2e4dfbd7d2acbdc3c5534dcf282d44fa94dde44b4b5ee9319b8","schema_version":"1.0","event_id":"sha256:2cf8798bc6ead2e4dfbd7d2acbdc3c5534dcf282d44fa94dde44b4b5ee9319b8"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/XZ6KU673RXJCDFKJNQGACJNLRV/bundle.json","state_url":"https://pith.science/pith/XZ6KU673RXJCDFKJNQGACJNLRV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/XZ6KU673RXJCDFKJNQGACJNLRV/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":"2026-05-27T00:31:20Z","links":{"resolver":"https://pith.science/pith/XZ6KU673RXJCDFKJNQGACJNLRV","bundle":"https://pith.science/pith/XZ6KU673RXJCDFKJNQGACJNLRV/bundle.json","state":"https://pith.science/pith/XZ6KU673RXJCDFKJNQGACJNLRV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/XZ6KU673RXJCDFKJNQGACJNLRV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:XZ6KU673RXJCDFKJNQGACJNLRV","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":"de6550001f65ea19f000f8d635df0f482bea715eade709c791379eef9256cdc5","cross_cats_sorted":["cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-08-28T09:31:09Z","title_canon_sha256":"2ba3f20877dadd522332bfbf50455cc66980575dd5bbc31e024cf536d29778ff"},"schema_version":"1.0","source":{"id":"2408.15664","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2408.15664","created_at":"2026-05-17T23:38:52Z"},{"alias_kind":"arxiv_version","alias_value":"2408.15664v1","created_at":"2026-05-17T23:38:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.15664","created_at":"2026-05-17T23:38:52Z"},{"alias_kind":"pith_short_12","alias_value":"XZ6KU673RXJC","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"XZ6KU673RXJCDFKJ","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"XZ6KU673","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:2cf8798bc6ead2e4dfbd7d2acbdc3c5534dcf282d44fa94dde44b4b5ee9319b8","target":"graph","created_at":"2026-05-17T23:38:52Z","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":"Loss-Free Balancing achieves both better performance and better load balance compared with traditional auxiliary-loss-controlled load balancing strategies."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That dynamically updating per-expert biases from recent load statistics will maintain balance across training without introducing instability or unintended routing dynamics."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Loss-Free Balancing keeps expert loads balanced in MoE models by dynamically adjusting routing-score biases based on recent usage, avoiding auxiliary-loss interference and yielding better performance."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Mixture-of-Experts models reach higher performance with load balancing that avoids auxiliary loss gradients."}],"snapshot_sha256":"c306008b14ccbd0298b01962067596270de956a35d5f098978a086983d20b104"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"ae8508b9f31a7ff7425541690bf1452fc90bde2e6779ab251d817b0d51cca3d1"},"paper":{"abstract_excerpt":"For Mixture-of-Experts (MoE) models, an unbalanced expert load will lead to routing collapse or increased computational overhead. Existing methods commonly employ an auxiliary loss to encourage load balance, but a large auxiliary loss will introduce non-negligible interference gradients into training and thus impair the model performance. In order to control load balance while not producing undesired gradients during training, we propose Loss-Free Balancing, featured by an auxiliary-loss-free load balancing strategy. To be specific, before the top-K routing decision, Loss-Free Balancing will f","authors_text":"Chenggang Zhao, Damai Dai, Huazuo Gao, Lean Wang, Xu Sun","cross_cats":["cs.CL"],"headline":"Mixture-of-Experts models reach higher performance with load balancing that avoids auxiliary loss gradients.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-08-28T09:31:09Z","title":"Auxiliary-Loss-Free Load Balancing Strategy for Mixture-of-Experts"},"references":{"count":32,"internal_anchors":6,"resolved_work":32,"sample":[{"cited_arxiv_id":"2401.06066","doi":"","is_internal_anchor":true,"ref_index":1,"title":"DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models","work_id":"a9888d6d-bf47-4324-9834-7cc12ac3a78c","year":2024},{"cited_arxiv_id":"2406.11931","doi":"","is_internal_anchor":true,"ref_index":2,"title":"DeepSeek-Coder-V2: Breaking the Barrier of Closed-Source Models in Code Intelligence","work_id":"713a39be-8c4e-4ee3-8671-11cf9379262f","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"William Fedus, Barret Zoph, and Noam M. Shazeer. Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity. J. Mach. Learn. Res., 23: 0 120:1--120:39, 2021. URL http","work_id":"2f29f1ca-bd51-4381-acc7-95e946cde35f","year":2021},{"cited_arxiv_id":"2006.16668","doi":"","is_internal_anchor":true,"ref_index":4,"title":"GShard: Scaling Giant Models with Conditional Computation and Automatic Sharding","work_id":"52b3c9a6-2a27-45a7-ba2b-ebe4b5bb5a5f","year":2006},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Sgdr: Stochastic gradient descent with warm restarts","work_id":"21d4473d-9c67-4464-b3ea-96588e1d01e1","year":2016}],"snapshot_sha256":"45ecc5e36e60cebbf268394ae511828f42a1187c8166048c5f0e7f95a94d51e0"},"source":{"id":"2408.15664","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-15T12:50:11.218526Z","id":"9ed35d07-9958-4ddc-9949-761036dfa418","model_set":{"reader":"grok-4.3"},"one_line_summary":"Loss-Free Balancing keeps expert loads balanced in MoE models by dynamically adjusting routing-score biases based on recent usage, avoiding auxiliary-loss interference and yielding better performance.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Mixture-of-Experts models reach higher performance with load balancing that avoids auxiliary loss gradients.","strongest_claim":"Loss-Free Balancing achieves both better performance and better load balance compared with traditional auxiliary-loss-controlled load balancing strategies.","weakest_assumption":"That dynamically updating per-expert biases from recent load statistics will maintain balance across training without introducing instability or unintended routing dynamics."}},"verdict_id":"9ed35d07-9958-4ddc-9949-761036dfa418"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:1762093ea14cc86b7cc32aeecf1835c4f30c04ba0667fbec5a5146f2fbcd5ff1","target":"record","created_at":"2026-05-17T23:38:52Z","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":"de6550001f65ea19f000f8d635df0f482bea715eade709c791379eef9256cdc5","cross_cats_sorted":["cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-08-28T09:31:09Z","title_canon_sha256":"2ba3f20877dadd522332bfbf50455cc66980575dd5bbc31e024cf536d29778ff"},"schema_version":"1.0","source":{"id":"2408.15664","kind":"arxiv","version":1}},"canonical_sha256":"be7caa7bfb8dd22195496c0c0125ab8d718011f0f205abb2232fbcaf12a94cf6","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"be7caa7bfb8dd22195496c0c0125ab8d718011f0f205abb2232fbcaf12a94cf6","first_computed_at":"2026-05-17T23:38:52.511754Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:52.511754Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"abHUXc0+nMOVci7mAwA8Ws1A0puaAk9b/wdZ9s1pIhqDSJV+b4lyfQpKD9+BIZ8Y4mimhyzkIYZThaggbRIrAg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:52.512176Z","signed_message":"canonical_sha256_bytes"},"source_id":"2408.15664","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1762093ea14cc86b7cc32aeecf1835c4f30c04ba0667fbec5a5146f2fbcd5ff1","sha256:2cf8798bc6ead2e4dfbd7d2acbdc3c5534dcf282d44fa94dde44b4b5ee9319b8"],"state_sha256":"322f0816ce6fd967ccf988990c6411e0781e499afd6b0bbb74df732faa664360"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qJpXPHHe6O4FV+901zpRf9Nlb46gXmpw6bkwvNbKHD6O4ZzVUcxx76znMqAbmt2DOgnkxaPE+i8PQ475YsOwCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-27T00:31:20.952879Z","bundle_sha256":"e78e516cfb6342723f118ba97303e032a8391a758c770d4b3c6de8799a2a5ac2"}}