{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:3OE23X4BS4QSGXYMNHS4FPSTYG","short_pith_number":"pith:3OE23X4B","schema_version":"1.0","canonical_sha256":"db89addf819721235f0c69e5c2be53c19e30a28c3fde433e77be490a1d56dd3f","source":{"kind":"arxiv","id":"2505.19797","version":3},"attestation_state":"computed","paper":{"title":"The Avengers: A Simple Recipe for Uniting Smaller Language Models to Challenge Proprietary Giants","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chenxu Wang, Daling Wang, Hao Li, Jia Xu, Lei Bai, Linyao Chen, Peng Ye, Qiaosheng Zhang, Shi Feng, Shuyue Hu, Wanli Ouyang, Xinrun Wang, Yiqun Zhang, Zhen Wang","submitted_at":"2025-05-26T10:29:42Z","abstract_excerpt":"Proprietary giants are increasingly dominating the race for ever-larger language models. Can open-source, smaller models remain competitive across a broad range of tasks? In this paper, we present the Avengers -- a simple recipe that leverages the collective intelligence of these smaller models. The Avengers builds upon four lightweight operations: (i) embedding: encode queries using a text embedding model; (ii) clustering: group queries based on their semantic similarity; (iii) scoring: scores each model's performance within each cluster; and (iv) voting: improve outputs via repeated sampling"},"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":"2505.19797","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-26T10:29:42Z","cross_cats_sorted":[],"title_canon_sha256":"a93a283b1c07076bdba4aebe3279af1ea2c2ede34a037432f8f3278872f78a50","abstract_canon_sha256":"c621503265b5c4e929832e96a2e5c6d10c71f0c85652e1aaae5f3be89367ed9b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:23:28.664925Z","signature_b64":"8adRujTV/oR2Y7cUh7IbzjlC8A6lvDIHkbKyw1tE4LiyYIexcsaJgeF5APtcvulOt/ZrR8+7NOSIGW6y6tCBAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"db89addf819721235f0c69e5c2be53c19e30a28c3fde433e77be490a1d56dd3f","last_reissued_at":"2026-07-05T11:23:28.664375Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:23:28.664375Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Avengers: A Simple Recipe for Uniting Smaller Language Models to Challenge Proprietary Giants","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chenxu Wang, Daling Wang, Hao Li, Jia Xu, Lei Bai, Linyao Chen, Peng Ye, Qiaosheng Zhang, Shi Feng, Shuyue Hu, Wanli Ouyang, Xinrun Wang, Yiqun Zhang, Zhen Wang","submitted_at":"2025-05-26T10:29:42Z","abstract_excerpt":"Proprietary giants are increasingly dominating the race for ever-larger language models. Can open-source, smaller models remain competitive across a broad range of tasks? In this paper, we present the Avengers -- a simple recipe that leverages the collective intelligence of these smaller models. The Avengers builds upon four lightweight operations: (i) embedding: encode queries using a text embedding model; (ii) clustering: group queries based on their semantic similarity; (iii) scoring: scores each model's performance within each cluster; and (iv) voting: improve outputs via repeated sampling"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.19797","kind":"arxiv","version":3},"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/2505.19797/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":"2505.19797","created_at":"2026-07-05T11:23:28.664435+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.19797v3","created_at":"2026-07-05T11:23:28.664435+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.19797","created_at":"2026-07-05T11:23:28.664435+00:00"},{"alias_kind":"pith_short_12","alias_value":"3OE23X4BS4QS","created_at":"2026-07-05T11:23:28.664435+00:00"},{"alias_kind":"pith_short_16","alias_value":"3OE23X4BS4QSGXYM","created_at":"2026-07-05T11:23:28.664435+00:00"},{"alias_kind":"pith_short_8","alias_value":"3OE23X4B","created_at":"2026-07-05T11:23:28.664435+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.22385","citing_title":"MetaPS: Adaptive Programmatic Strategy Selection for Market Agents","ref_index":138,"is_internal_anchor":false},{"citing_arxiv_id":"2606.18774","citing_title":"RouteJudge: An Open Platform for Reproducible and Preference-Aware LLM Routing","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2606.04378","citing_title":"DLLG: Dynamic Logit-Level Gating of LLM Experts","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23530","citing_title":"MTRouter: Cost-Aware Multi-Turn LLM Routing with History-Model Joint Embeddings","ref_index":2,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3OE23X4BS4QSGXYMNHS4FPSTYG","json":"https://pith.science/pith/3OE23X4BS4QSGXYMNHS4FPSTYG.json","graph_json":"https://pith.science/api/pith-number/3OE23X4BS4QSGXYMNHS4FPSTYG/graph.json","events_json":"https://pith.science/api/pith-number/3OE23X4BS4QSGXYMNHS4FPSTYG/events.json","paper":"https://pith.science/paper/3OE23X4B"},"agent_actions":{"view_html":"https://pith.science/pith/3OE23X4BS4QSGXYMNHS4FPSTYG","download_json":"https://pith.science/pith/3OE23X4BS4QSGXYMNHS4FPSTYG.json","view_paper":"https://pith.science/paper/3OE23X4B","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.19797&json=true","fetch_graph":"https://pith.science/api/pith-number/3OE23X4BS4QSGXYMNHS4FPSTYG/graph.json","fetch_events":"https://pith.science/api/pith-number/3OE23X4BS4QSGXYMNHS4FPSTYG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3OE23X4BS4QSGXYMNHS4FPSTYG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3OE23X4BS4QSGXYMNHS4FPSTYG/action/storage_attestation","attest_author":"https://pith.science/pith/3OE23X4BS4QSGXYMNHS4FPSTYG/action/author_attestation","sign_citation":"https://pith.science/pith/3OE23X4BS4QSGXYMNHS4FPSTYG/action/citation_signature","submit_replication":"https://pith.science/pith/3OE23X4BS4QSGXYMNHS4FPSTYG/action/replication_record"}},"created_at":"2026-07-05T11:23:28.664435+00:00","updated_at":"2026-07-05T11:23:28.664435+00:00"}