{"paper":{"title":"Towards Self-Evolving Agentic Literature Retrieval","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"PaSaMaster turns literature retrieval into a self-evolving process that ranks papers by relevance without generating sources, outperforming GPT-5.2 by 30% at 1% cost with zero hallucinations.","cross_cats":[],"primary_cat":"cs.IR","authors_text":"Fenyi Liu, Jing Kang, Jingyi Chai, Siheng Chen, Sikai Yao, Tian Jin, Tingjia Miao, Wenhao Wang, Xianghe Pang, Yuwen Du, Yuzhi Zhang","submitted_at":"2026-05-14T03:17:31Z","abstract_excerpt":"As large language models reshape scientific research, literature retrieval faces a twofold challenge: ensuring source authenticity while maintaining a deep comprehension of academic search intents. While reliable, traditional keyword-centric search fails to capture complex research intents. Frontier LLMs can handle complex research intents, but their high cost and tendency to hallucinate remain key limitations. Here we introduce PaSaMaster, a self-evolving agentic literature retrieval system that produces relevance-scored paper rankings with evidence-grounded recommendations through iterative "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"PaSaMaster outperforms GPT-5.2 by 30.0% at a mere 1% of the computational cost while ensuring zero source hallucination, and improves F1-score by 15.6X over traditional keyword retrieval on the PaSaMaster Benchmark across 38 disciplines.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The PaSaMaster Benchmark faithfully represents real-world scientific search intents and that the iterative evidence-ranking process reliably improves results without introducing selection bias or new failure modes.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"PaSaMaster is a self-evolving agentic literature retrieval system that improves F1-score by 15.6X over keyword search and outperforms GPT-5.2 by 30% at 1% cost with zero source hallucination across 38 disciplines.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"PaSaMaster turns literature retrieval into a self-evolving process that ranks papers by relevance without generating sources, outperforming GPT-5.2 by 30% at 1% cost with zero hallucinations.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"fe8c948871016e9bc364c8d563f7c8821b3374f3e203eef78267b169e9f9a559"},"source":{"id":"2605.14306","kind":"arxiv","version":1},"verdict":{"id":"84c8003a-4a00-409a-bd0d-2046a6969954","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T02:33:11.578720Z","strongest_claim":"PaSaMaster outperforms GPT-5.2 by 30.0% at a mere 1% of the computational cost while ensuring zero source hallucination, and improves F1-score by 15.6X over traditional keyword retrieval on the PaSaMaster Benchmark across 38 disciplines.","one_line_summary":"PaSaMaster is a self-evolving agentic literature retrieval system that improves F1-score by 15.6X over keyword search and outperforms GPT-5.2 by 30% at 1% cost with zero source hallucination across 38 disciplines.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The PaSaMaster Benchmark faithfully represents real-world scientific search intents and that the iterative evidence-ranking process reliably improves results without introducing selection bias or new failure modes.","pith_extraction_headline":"PaSaMaster turns literature retrieval into a self-evolving process that ranks papers by relevance without generating sources, outperforming GPT-5.2 by 30% at 1% cost with zero hallucinations."},"references":{"count":31,"sample":[{"doi":"","year":2025,"title":"PaSa: An LLM Agent for Comprehensive Academic Paper Search , author=. 2025 , eprint=","work_id":"f302f560-723c-47fd-bf5b-f3af7936ea37","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks , author=. 2021 , eprint=","work_id":"4696b88f-1bab-4fde-ba0a-dd8d18a3326d","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Siren's Song in the AI Ocean: A Survey on Hallucination in Large Language Models , author=. 2025 , eprint=","work_id":"cbb71f3c-48bc-447a-81cc-95c936c464e7","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2019,"title":"BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding , author=. 2019 , eprint=","work_id":"4c41a8dc-7f3a-4093-96ed-9cb29b926433","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","ref_index":5,"cited_arxiv_id":"2303.08774","is_internal_anchor":true}],"resolved_work":31,"snapshot_sha256":"fb15c1405d0fb70249914a3c2ead6dbd247b3b04b61f63fc3340d78d57d8dcd7","internal_anchors":1},"formal_canon":{"evidence_count":2,"snapshot_sha256":"839200b48fc54b714e0d0e09d4f38cf8e023b93a9412625d8298b794002232ca"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}