{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:XJ6SSXAM5BS3AV76FOZHVMGKIN","short_pith_number":"pith:XJ6SSXAM","schema_version":"1.0","canonical_sha256":"ba7d295c0ce865b057fe2bb27ab0ca43731e7ab25ba6294017e2f32648297a7b","source":{"kind":"arxiv","id":"2501.17858","version":2},"attestation_state":"computed","paper":{"title":"Improving Your Model Ranking on Chatbot Arena by Vote Rigging","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chao Du, Minhao Cheng, Min Lin, Qian Liu, Rui Min, Tianyu Pang","submitted_at":"2025-01-29T18:57:29Z","abstract_excerpt":"Chatbot Arena is a popular platform for evaluating LLMs by pairwise battles, where users vote for their preferred response from two randomly sampled anonymous models. While Chatbot Arena is widely regarded as a reliable LLM ranking leaderboard, we show that crowdsourced voting can be rigged to improve (or decrease) the ranking of a target model $m_{t}$. We first introduce a straightforward target-only rigging strategy that focuses on new battles involving $m_{t}$, identifying it via watermarking or a binary classifier, and exclusively voting for $m_{t}$ wins. However, this strategy is practica"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2501.17858","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-29T18:57:29Z","cross_cats_sorted":["cs.AI","cs.CR","cs.LG"],"title_canon_sha256":"c25bb5653408a77f8e96c5a01f0f311dc9d83ca6d596f01324cb541dcfad3809","abstract_canon_sha256":"734f9f42184691acc88ff15baed9553cf34a7ba6562e53d083bbcfafeef04ccf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:51:21.321935Z","signature_b64":"Qt1rOQUNKzp179IUc1WeOrYUWmXX3aOxaPFOTZLAa14TgFF/cGdLVORVs+JfSoz/pW0pXgTppFxa7qKOPk3fDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ba7d295c0ce865b057fe2bb27ab0ca43731e7ab25ba6294017e2f32648297a7b","last_reissued_at":"2026-07-05T11:51:21.321411Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:51:21.321411Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improving Your Model Ranking on Chatbot Arena by Vote Rigging","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chao Du, Minhao Cheng, Min Lin, Qian Liu, Rui Min, Tianyu Pang","submitted_at":"2025-01-29T18:57:29Z","abstract_excerpt":"Chatbot Arena is a popular platform for evaluating LLMs by pairwise battles, where users vote for their preferred response from two randomly sampled anonymous models. While Chatbot Arena is widely regarded as a reliable LLM ranking leaderboard, we show that crowdsourced voting can be rigged to improve (or decrease) the ranking of a target model $m_{t}$. We first introduce a straightforward target-only rigging strategy that focuses on new battles involving $m_{t}$, identifying it via watermarking or a binary classifier, and exclusively voting for $m_{t}$ wins. However, this strategy is practica"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.17858","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/2501.17858/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2501.17858","created_at":"2026-07-05T11:51:21.321477+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.17858v2","created_at":"2026-07-05T11:51:21.321477+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.17858","created_at":"2026-07-05T11:51:21.321477+00:00"},{"alias_kind":"pith_short_12","alias_value":"XJ6SSXAM5BS3","created_at":"2026-07-05T11:51:21.321477+00:00"},{"alias_kind":"pith_short_16","alias_value":"XJ6SSXAM5BS3AV76","created_at":"2026-07-05T11:51:21.321477+00:00"},{"alias_kind":"pith_short_8","alias_value":"XJ6SSXAM","created_at":"2026-07-05T11:51:21.321477+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2510.18333","citing_title":"Position: LLM Watermarking Should Align Stakeholders' Incentives for Practical Adoption","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10851","citing_title":"The Generalized Turing Test: A Foundation for Comparing Intelligence","ref_index":14,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XJ6SSXAM5BS3AV76FOZHVMGKIN","json":"https://pith.science/pith/XJ6SSXAM5BS3AV76FOZHVMGKIN.json","graph_json":"https://pith.science/api/pith-number/XJ6SSXAM5BS3AV76FOZHVMGKIN/graph.json","events_json":"https://pith.science/api/pith-number/XJ6SSXAM5BS3AV76FOZHVMGKIN/events.json","paper":"https://pith.science/paper/XJ6SSXAM"},"agent_actions":{"view_html":"https://pith.science/pith/XJ6SSXAM5BS3AV76FOZHVMGKIN","download_json":"https://pith.science/pith/XJ6SSXAM5BS3AV76FOZHVMGKIN.json","view_paper":"https://pith.science/paper/XJ6SSXAM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.17858&json=true","fetch_graph":"https://pith.science/api/pith-number/XJ6SSXAM5BS3AV76FOZHVMGKIN/graph.json","fetch_events":"https://pith.science/api/pith-number/XJ6SSXAM5BS3AV76FOZHVMGKIN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XJ6SSXAM5BS3AV76FOZHVMGKIN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XJ6SSXAM5BS3AV76FOZHVMGKIN/action/storage_attestation","attest_author":"https://pith.science/pith/XJ6SSXAM5BS3AV76FOZHVMGKIN/action/author_attestation","sign_citation":"https://pith.science/pith/XJ6SSXAM5BS3AV76FOZHVMGKIN/action/citation_signature","submit_replication":"https://pith.science/pith/XJ6SSXAM5BS3AV76FOZHVMGKIN/action/replication_record"}},"created_at":"2026-07-05T11:51:21.321477+00:00","updated_at":"2026-07-05T11:51:21.321477+00:00"}